WeRelate talk:Source Committee/Archive

Tasks

Dallan has given the committee some ideas or talking points to start the ball rolling. His comments have been copied from WeRelate_talk:Sources:

• The plan is to move all of the sources over to the new format automatically, so we don't need to move anything over manually right now, although there will probably be some clean-up work afterward.

• In order for sources to be found most easily when someone starts typing source titles in data input fields or when they search for them, we need to come up with a small set of simple rules and vocabularies for people to use when creating source titles. Ideally the title format would be such that once I've typed the first few words of the title into an input field, I can browse a reasonably-small list of titles that start with those words and find the source that I'm looking for. We have some discussion about title format on this page and on WeRelate talk:Source page titles, but I'm not sure it adequately covers all types of sources. It would be great if the committee could develop this set of rules that covers most (if not all) types of sources.

• It would also be great if the committee could come up with a list of additional fields that we're not tracking (like location/repository) that we need to track, and perhaps some suggested templates. For example, I think that a repeatable "location" field, where the user would first select the type of location (worldcat, a library, an archive, a website, etc.) and then enter a way to find the source in that location (an ISBN number, a catalog number, a URL) would be nice. We could have the ISBN number link to online booksellers or worldcat, a library/archive link to its Repository page, etc. And since we're going to start creating Repository: records, it would be nice to know what fields we want to track for repositories.

• I'd like to be peripherally involved in this at the beginning. I've got a lot of usability enhancements that I'd like to put in place before the Federation of Genealogical Societies conference mid-August. If the committee could do these things between now and mid-August, after FGS is over I can start writing the program to move sources over to the new format. Then there will be some clean-up work, but we should be able to automate most of it.

--Dallan

Types of Sources

One of the things I think we need to quickly reach a decision on is the TYPES of sources we are going to categorize. Wrhelp has already done a great job of listing the various types on WeRelate_talk:Source_page_titles. Once we settle on the TYPES, then we should decide on how each TYPE should be formatted for the source title (again, Wrhelp has outlined most of this, we just need to reach a decision). Then decide on templates and finally, what fields should be "tracked" in the search results as Dallan mentioned above.--Ronni 01:49, 14 July 2007 (MDT)

What Elizabeth Shown Mills Says about Source Types

From: Skillbuilding: Citing Your Sources Added by Jillaine 10:05, 15 July 2007 (MDT)

Book:

• name of author
• title of book (in italics or underlined)
• place of publication
• name of publisher
• year of publication
• page (and possibly volume) number

Periodical article:

• name of author
• title of article (in quotation marks)
• title of periodical (in italics or underlined)
• volume number
• date of publication (month and year)
• page number

Original record:

• title of document (i.e. Deed: John Brown to Sam Smith)
• date(s) written and/or recorded
• collection name (e.g.: Probate Judge Files; or Thomas Tidbury Collection)
• book/page of document (if recorded in bound manuscript volumes)
• file or box and document number
• repository name
• repository location (city and state suffice for public institutions; if document is in private possession, give street address.)

Jillaine, thank you. Now we're looking at things that might be in our template for various types of sources, once we've settled on certain types and in what format to name their pages. What boxes do we want to have in editing the Source Page for folks to fill out?

The current one-size-fits-all source template has some nice features I wouldn't want to lose: boxes for surnames covered, places covered, and year range. These are not necessary when you're writing up your research, but in a wiki another purpose of source listings is to help searchers find relevant sources.

Book template: obviously for our wiki purposes the title has its own box, and doesn't need to be entered in italics. I like the idea of having separate boxes for place of publication, date of publication, and publisher. That saves me having to think (as we do at present), oh dear, what counts as "publication information"? Obviously the page number should go in your citation, not on the Source Page.

As discussed elsewhere, another wiki purpose is to let people know all the different ways they can get ahold of this book (or any other authored document). This gets complicated, as there are types of locations (libraries, online databases) and multiple locations within each. (Dallan has discussed this elsewhere WeRelate talk:Sources8.3.) And since we'll have a new namespace for repositories, none of these templates needs to include repository locations.

Same thoughts on the periodical template: no need for itals or quotes since article and journal title are in their own template boxes, and the actual page numbers belong in the citation itself, not on the Source Page.

I'm not sure the "original record" is a helpful starting template for us, since it seems to assume that every individual record would have its own Source Page. I don't think we could cope with having every single individual's WWI draft card having its own Source Page!

--Hh219 11:04, 15 July 2007 (MDT)

Discusssion of Source Types

Here is a rough draft of the TYPES as I see it. Add to, delete and/or combine: --Ronni 01:49, 14 July 2007 (MDT)

• Books (would include sources that have a clearly defined author, i.e., books, articles, manuscripts, biographies, research guides, etc)

• Periodicals (magazines, newspapers, newsletters, and journals)

• Census records

• Vital record types (including land records, tax records, church records, etc)

• Websites

• Message boards

• Audio (from Jillaine's list below)

• Video (from Jillaine's list below)

• Microfilm/Microfiche (from Jillaine's list below)

I think one of our goals is to limit the number of types, but yet at the same time, make it so a type will be so clearly defined that it will be relatively easy for a user to add their source to the index into one of these types. --Ronni 01:49, 14 July 2007 (MDT)

We might want to look at most common genealogy programs to see what source type definitions they use. I use Family Treemaker, and its source types are:

• Ancestry.com
• Audio
• Book
• Card
• Census
• Church Record
• Civil Registry
• Electronic
• Family Archive CD
• Genealogy.com Data CD
• Internet
• Letter
• Magazine
• Manuscript
• Map
• Microfiche
• Microfilm
• Newspaper
• Official Document
• Other
• Photograph
• Tombstone
• Unknown
• Video

Not saying that we should use all of these, but that's what a lot of people will be familiar with and what may also be imported with GEDCOMs. Jillaine 06:33, 14 July 2007 (MDT)

I'm wondering if we're re-inventing the wheel a bit here? Over at WeRelate_talk:Source_page_titles Dallan made a rough draft of types of sources which basically divided them into 3 types:

• those with authors (such as most books and manuscripts). Their source pages would be named "author surname, author first name, title".

• those without authors but with titles (such as newspapers and periodicals). Their source pages would be named, "title".

• those with neither authors nor titles but associated with places (such as vital records, cemeteries). Their source pages would be named "country, state, county, place".

I like this because it seems pretty straightforward. Two questions: (1) is this a good way to divide things up very generally? and, (2) is this exhaustive? What sources don't fall into any of these three bins? (Dallan also had a category for websites and message boards that contain original information, but I'm wondering if they can't be sorted into these three bins.)

Assuming that this setup or something similar is appropriate to start with, then we would need to think about what templates we need *within* each bin.--Hh219 12:53, 14 July 2007 (MDT)

Creating a source page will be a two-step process for the user. First they are confronted with how to format the title of the source page. This process needs to have just a few, very simple "rules" to follow. Once the user gets to the template portion of setup, then they will be confronted with the finer details of their source, where I actually think the bulk of our discussion has been focused. Hh291 has reminded us that most of these "rules" have already been laid out for us at WeRelate_talk:Source_page_titles. While reading the discussion, I think I've come to the understanding that the TYPE of source is not initially that important when contructing the source title. It's how they answer the following questions as to how the title gets formatted: --Ronni 01:02, 16 July 2007 (MDT)

(1) Does this source have a clearly distinct author AND title?

Format: <Author (last name, first name). Title>

(NOTE: a "period" is used instead of a "comma" between author and title)

(2) Does this source have a title, but no distinct author?

Format: <Title>

(3) Does this source have neither a distinct title or author?

I actually see #3 as being subdivided into asking whether its a CENSUS or a WEBSITE or a MESSAGE FORUM (and other types I'm sure) as I see slightly different formatting problems for each, one being the number of characters we are limited to. I believe it was once suggested that websites have the URL included in the title. I don't believe that would work because of the length of some of the URLs.

When I refer to "asking" the user a series of questions, I wasn't meaning that literally, but now I wonder if source page creation can not be designed like a flowchart type thing? A setup wizard perhaps? --Ronni 01:02, 16 July 2007 (MDT)

Repositories [4 December 2007]

Aside from the sources, we also need to be thinking of templates for repositories. As it is now, repositories are included in the source index. The goal is to make "repositories" a new namespace.

Dallan suggested for starters that the fields to be filled in for a repository would be address, phone, URL, and of course a description field. Are there other things?

Call me crazy, but I'm actually excited about having repository pages. They can be the go-to place when you're headed out to Newark, Ohio, or to Montpelier, Vermont, and you want to know the straight skinny about where you're going, all the way from where to park and eat and walk the dog, to whether they use Dewey Decimal or Library of Congress. They can be where you can tell about the quirky local resources that are there and nowhere else, that don't always get on the libraries' web pages.

Are ancestry and heritagequest repositories? How about the various genealogical society websites with useful information?

--Hh219 19:54, 14 July 2007 (MDT)

We'll have to agree on the definition of what a repository is before we can decide on what constitutes a repository. Calling upon my own definition, I believe Ancestry.com and Heritage Quest are repositories. But what about a family website? Is it a source or a repository? --Ronni 02:22, 16 July 2007 (MDT)

The thing that really identifies a "source" of information is that you or others can go back to it, to be certain that it says what you think it says (verification). If you find something on Ancestry, say in one of its historical data bases, you can indeed go back to it, more or less anytime you want, more or less indefintely...Assuming Ancestry continues in operation in one form or another. In theory, you should go for the original, but in practice, using Ancestry as a surrogate for the original seems reasonable---but probaby what you need to do is cite the original source with the "fide Ancestry....tag". But if you can't go back to the original, its not a useful source of information.

And that's what makes things like personal web pages, or GEDCOM's on Ancestry, etc., virtually useless as sources. Information on such "sources" is ephemeral at best, subject to constant change. If you use it as a source, you can almost be certain its going to dissappear on you at some point in the future. Then you can't verify what it had to say, so citing it as a source is meaningless---unless, of course, they provided the citation to THEIR source of information, which is what you'd need to follow up on---not the personal web site. And again, what probably would work best is citing the original source (assuming they give one), with the "fide tag" to indicate you are taking it on faith. That way, if all else fails, it can still be followed up.

On the other hand, if all you have is someone's personal notation, and they don't tell you where they got the information from, or its of such a nature that it can not be followed up, there's probably no point in eveing sourcing it Bill 17:18, 4 December 2007 (EST)

My sense is that a repository contains sources, so ancestry and heritage would qualify. Family websites might provide new information themselves, or contain, say, transcriptions or excerpts from elsewhere, so they could actually be both at the same time. But when I'm doing data entry I don't want to have to think that hard. So I'd rather treat them as sources. Would that create any problems, really, aside from being a bit impure? --Hh219 09:02, 16 July 2007 (MDT)

Thinking about Repository pages intermingles with my thinking about Source pages. Here is my thinking so far:

The Data (birth date, picture, family story) - a fact, picture or story about a person

The Source (book, gravestone, 2nd Cousin, online transcription) - a physical thing or person, including WWW, where you found the Data

The Repository (WWW, email, CD, cemetery, 2nd Cousin's details) - the physical place in which the Source resides.

So what do you think about this scenario? I have my photocopy from a microfilm copy of a death record book, the microfilm being kept at IRAD (a depository). My guess would be that Data is the date and details, Source is the actual book, Type is microfilm copy, and Respository is the place where the microfilm lies (the repository). --Debbie V. 11:47, 22 July 2007 (MDT)

I think I agree, Debbie, although it seems like cemeteries want to be sources sometimes, places sometimes, and repositories sometimes, and it seems like rather much to require triple inputting of every cemetery on the planet! In your example, I'm inclined to think of the microfilm as identical to the original, so that this source might have several repositories -- IRAD, FHL, and wherever the county death record book itself is these days.

--Hh219 12:47, 22 July 2007 (MDT)

I think we are all in agreement that a Repository is a place, with a geographic or internet address. Just like the Sources, we may want to decide how local we want the Repositories to be. For example, within the NARA there are many "addresses" where Sources are "housed" both in Washington D.C., regional centers, and online. Do we just want a general NARA repository or do we want to have a Repository for different "subdivisions" of NARA?

Hh219, yes, the microfilm and actual record book would be the same Source, but different types or forms within that Source.--Debbie V. 15:35, 22 July 2007 (MDT)

Discussion [7 August 2007]

Ok everyone. I went ahead and created this page (came up with ever so original title of "Source Committee" all on my own too :)) so we could have a place to "meet." Feel free to move it or re-title it to whatever is more appropriate. I figured we could use the project page for our more "finalized" ideas and use the Talk page for discussion. --Ronni 11:40, 12 July 2007 (MDT)

Thanks, Ronni. Jumping right in from Dallan's comments, we need to somehow be clear that the source title is not *necessarily* the same as the title of the book or magazine article or whatever. For instance, take an easy example, "The History of Ancient Windsor [CT]" by Henry R. Stiles. The source title, at least as we've been thinking, would be: "Stiles, Henry. History of Ancient Windsor." The book's title -- what you would put in the "title" box when entering it -- is either "The History of Ancient Windsor" or "History of Ancient Windsor" depending on the convention we choose.

I have liked the idea of doing source title this way, with author first, but having looked at it now, I wonder if maybe this is just too arcane and confusing for newbies and we should not try to maintain this distinction, just have source title and title box contents always be the same. What do people think?

I don't think the distinction will be too much of a stumbling block, especially if we do a "source manual" to help users when adding sources. Author first, especially with book types, is the way to go, in my opinion. I have created many book sources with author first and it has been a breeze to call up the source in the input field on Person and Family pages. Remember, the first few words of the title has to be correct for the title you are searching for, or it won't be called up in the list. It is much easier to simply type in that last name than trying to remember what are the first few words of the book. --Ronni 02:03, 14 July 2007 (MDT)

This example raises other questions, such as, what should we do with initial "the," "an," and "a"? (I say ditch 'em but it's hard to remember.) Should different editions be listed separately, and if so how? (Stiles 1859 is way different from Stiles 1892!) Should we treat a two-volume set, which Stiles is, as one source? Is there a professional librarian lurking who'd like to help?

--Hh219 13:21, 12 July 2007 (MDT)

There is a professional librarian "lurking," but we at Allen County Public Library are so engrossed in the imminent FGS meeting that I simply do not have the time right now to devote to this important subject. Again, I believe any decisions and the bulk of future discussion should be held off until Elizabeth Shown Miller's new edition of Evidence becomes available at the meeting, then give the matter more time and thoughtful consideration.

Mary--Kittydoc 08:44, 7 August 2007 (MDT)

If we go with author first, then title, I don't think "the," "an", etc will make much of a difference. Now, as to listing different editions on the same page, I'm voting for one page for all editions, even realizing the editions might be vastly different from each other. These differences, of course, should be noted on the source page. With all versions of a book, for example, on the same page, there won't the issue of how to make a source title for each edition. Also, one edition might be online while the other is not. Searching for the source in WeRelate, I would have to visit each page to see the various editions whereas I would personally like to click once and see everything on one page: the different editions, volumes, whether its online or not, etc. --Ronni 02:03, 14 July 2007 (MDT)

From an end-user perspective, what is it that they're most likely to a) use, and b) search by? Seems to me like Title more likely to come first. What does that woman author say-- the one who wrote about recommended citations for genealogy? (I've got her book somewhere... let me go look...) Jillaine 06:35, 14 July 2007 (MDT)

What I think of are those annoying similar sounding books that can only be distinguished by the author - i.e. The English Genealogical History of Henry Adams vs. The Somerset Ancestry of Henry Adams. Whoever is using/looking for the source probably knows the author and will want to use that to make sure they have the right record. For search purposes, having the author field filled in (which it isn't currently on most records), would greatly help search. But we still have the question of how the source should display in the search results (a question that will have to wait until the new search) and on person pages as a source when someone doesn't do something fancy. In both contexts, I think I vote for having the author visible. I think it would look better and be more intelligible as a source on person pages, and it would be easier to enter it in the source box when making one-off changes.--Amelia.Gerlicher 18:15, 14 July 2007 (MDT)

Multiple forms of same source

Here's an sourcing issue that I have wondered about and may relate to our Source index - a source that is available in several "forms". For example, last week I requested a photocopy of a marriage record from IRAD (Illinois Regional Archives Depository) at SIU (Southern Illinois University). I'm not sure if they archive the actual record books (birth and death registers, probate papers, etc) or if they just have microfilmed copies. What they sent me was the print out from a microfilm copy of the page of the actual handwritten marriage register from 1828, St. Clair County, Illinois. My question is do I choose my source as the microfilm or the actual record book?

In some cases, a particular record, such as Book A of the death records of a particular county can be available in it's original form (at the court house), a microfilm (an archive or library), an online database, a transcribed book, or a website transcription.

When we create the Source pages, will we have a separate Source for each of these forms?

So we have to decide (1) how to list sources with different editions (i.e., one page vs different pages) and (2) how to list sources that are essentially the same, but appear as different TYPES. Debbie's example could also apply to a book that appears in hard copy or on CD. One page vs two pages?

If the template design can handle inputting different types on one page, that's something to consider. My personal opinion is to have one page with all the differing information on it, that is, different editions and types all included, but differentiated on the same page. I'm not sure how a template design would look though. It may be too confusing. --Ronni 11:52, 14 July 2007 (MDT)

I'm with Ronni. We should not cater to the confused idea that a CD or a microfilm is a "source" when it's a copy of an underlying book or vital record. The source is book A of the death records of St. Clair County, Illinois. The IRAD microfilm is the type. The FHL microfilm is another type. The original (if you went to the IRAD office) is another type. Buuuuuut....I would be inclined to say that a published transcription of the death records would be a different source (because of the many wonderful possibilities for error that don't exist in the process of microfilming). Here's where I would like to read what Ms. Mills would say?--Hh219 12:13, 14 July 2007 (MDT)

I'm mostly with you both. But, I think that as far as the transcription versus the original, that they need to be the same source/page. One of the purposes of the source page, at least as far as I see it, is information on the source, where it can be found, and what the strengths/weaknesses of each version are. Do we want to have 4 versions of every census? The federal microfilm, the ancestry.com images, the heritagequest images, and the genweb text translation? That seems silly. Much more useful is one page and where it can be found, and a note that the text transcription found at X is good/bad/badly formatted, etc. It's got to be up to the person adding the source to a page to note which version they used (like they will have to note page number, etc.) and we in turn can maintain a useful single page on each source, at least as far as we can manage.--Amelia.Gerlicher 18:15, 14 July 2007 (MDT)

With censuses, "United States, Illinois, St. Clair County, 1900" would be the source name for the actual census images, whether they be on federal microfilms seen at the library, or on ancestry or heritagequest. If anyone wants to cite John Doe's laborious 1955 transcription of the census, that should be a separate citation, something like "Doe, John. Transcription of 1900 Census, St. Clair County, Illinois" if that was his title. The beauty of this is that we're on the internet now, and we can and should link to various transcriptions from the main census page. Transcriptions are fundamentally different from actual images. I'm willing to bet Ms. Mills will take a stand on that.

Another way of looking at it is, a source citation says "This is what I looked at." If I looked at the original census, I want my sourcing to be able to say that, and not be lumped in with someone who looked only at the transcription.--Hh219 19:46, 14 July 2007 (MDT)

I'm going to complicate my own analysis and say that sources like this raise additional issues: Source:Ricker, Jacquelyn Ladd. The Ricker Compilation of Vital Records of Early Connecticut - that is, sources, like many CD compilations out there, that combine data from multiple sources in what purport to be close to the original format. Even though this is arguably just another type that should be combined with the Barbour VR page, I think it should have a separate page because the merits/usefulness/details for this particular source are worthy of a discussion separate from the Barbour collection. I think that distinction holds even if you were to tell me that the Ricker CD contains pdfs of the actual Barbour books, but I do not think I'd be in favor of another page if the cd contained *only* Barbour collection data. There's a fuzzy line there, because I would want to avoid creating a different source page for every version of the Windsor, CT Barbour collection vital records out there (for example) and we're going to need a rule about when that's appropriate, a rule that someone sourcing info from their database from a source they saw five years ago can use.

Here too I think the operative question is, what did you look at? If you went to the Connecticut State Library and saw the Barbour Collection cards themselves, that should be differently citable than going to your own library and reading the published book. Likewise I have no trouble with Ricker's compilation being cited as what it is, if that's what you looked at. Again, linkage is all. We can make links and compare without having to lump all our sourcings together.--Hh219 19:46, 14 July 2007 (MDT)

Hmmm, I think we have a disagreement that's pretty fundamental. And maybe I'm reading it wrong, but my vision of what we're doing has me agreeing with what hh219 just said but thinking that is has no bearing on what source pages should be. I view the question of what one should cite (where the "what I looked at" question is relevant) as almost entirely separate from the title and contents of a source page on WeRelate, which is independent of repository or edition and thus independent of the 'what I looked at' dilemma. What I cite (theoretically) includes things like the edition, volume, page, repository, url, etc. that may change depending on where and when I looked at a source. But the source itself, regardless of where I looked at it, is what gets its own page. What and how people should link to that page when sourcing info on a person/family page might influence how we title it, but I view the link to the source page as just one part of the source citation that should include other information.

Perhaps we should back up and talk about the goals of source pages. I have been operating under a theory that they are best for 1) collecting and identifying repositories, webpages, etc. where a source can be located; 2) providing a place to describe and discuss the content, usefulness and weaknesses of the source; and 3) linking the source to localities and surnames, to aid searching and research. When people see a source on a person/family page, the link should usually be just part of it, and clicking on it tells someone more about the source. But there on the person/family page is the citation information for that particular bit of info. This is why I'm not excited about separate pages for transcriptions vs. originals or articles within periodicals, although I'm willing to be convinced, because I think they clutter up the search and the project unnecessarily by distributing information to a number of pages that can be collected more usefully in one place.--Amelia.Gerlicher 22:28, 14 July 2007 (MDT)

Good points. In the genealogy program I use, there's the "source" which has its own little page, and there are "citation details" that you hopefully enter every time you use the source. In the easy case, the "source" would be a book title and author and publisher, and the "citation detail" would be the page number. In a hard case like censuses, some people choose to make each household a separate source (which drives me crazy and leaves very little to go in the citation detail), some make each county and each year a separate source (my preference, but it's kind of arbitrary), and some I gather make the entire US census for a given year the source (which makes the citation detail huge).

I think we're arguing about the same kind of judgment call along a spectrum here. We can kind of aggregate sources -- have a source page that include all versions of the Barbour collection, for instance -- and distinguish which version we looked at whenever we cite that general source. That keeps the source page more comprehensive, and I think Amelia is convincing me that in a wiki it may be better in general to lean toward aggregating versions of a source on a single page, more important than in one's own program. But there is a price for doing so: it does put more of a burden on the citation detail. If I get lazy and neglect to put any citation detail in (or if I did the research before I ever tried to think about this stuff!), and just cite "Barbour" in general, then the next reader won't be too well informed as to just what I looked at and what weight to give it. But it's a wiki, so the next person along can improve it if they know which version they saw on that particular person.

So I might even agree to include transcriptions and originals on the same "source" page, although I still want to see what Mills says. But wait -- what is the "source" exactly? Many of us can and do rest content with seeing what's in the Barbour books. If that's problematic, we can refer to the cards themselves at CSL or on microfilm. But the best possible evidence for the birth of my ancestor Ezra Mead in Greenwich CT is none of these, it's the wonderful entry on page 48 of the first volume of the Greenwich commonplace book: "Timothy mead had a son born october ye 9 day 1737 and he called his name Ezra made."

In a profound sense, THIS is the source. And therefore, in theory, it could be encompassed within a gigantic "source" including Barbour and other transcriptions of Connecticut vital records and all the original vital records of every Connecticut town.

Um, actually, at least how I do documentation of my sources, the source title for this would be "Greenwich commonplace book, volume I" and would include its author and physical location. It's completely different than Barbour in my mind! (Apologies if I'm returning to this conversation after some page has been turned, but felt I had to respond.) Jillaine 12:18, 20 July 2007 (MDT)

I think that would be going too far, if only because it would take a long time to convince most people of it or even explain why we were doing it that way. Right now Amelia has moved me toward greater source aggregation -- all versions of Barbour on a single source page, with citation details saying which one you consulted and which page of which book it was in or whatever -- but not all the way. I would prefer to keep Greenwich Vital Records a separate source from that.

I hope this discussion will help us figure out how & why we make these decisions. I'm pretty sure there are ramifications I'm not picking up on.--Hh219 07:02, 15 July 2007 (MDT)

I'm all for combining different editions and volumes on the same page, but in order to do this they need to have very similar titles throughout all the editions/volumes/versions, etc AND by the same author. Amelia gave the example of The Ricker CD and being in favor of it having a separate page. I agree with her, but for a different reason and that being simply because it has a different TITLE and a different AUTHOR than the Barbour Collection. How can we create a citation on our Person/Family pages when the source input field could have a completely different title for the source we're referencing? Imagine if Ricker did ONLY have the Barbour Collection on the CD, then the argument was for them to possibly be combined on the same page because they in essence are the same source, in which case, how would you cite Ricker if you are linking to a source title that has Barbour on it? --Ronni 01:56, 17 July 2007 (MDT)

Really Broad/Internet/Periodical Sources

Thinking about this also raised (for me at least) a couple other side issues that can go somewhere else, but may still be under the committee's jurisdiction:

• Are we going to develop a set of rules for citing sources? For example, say I have seen the Barbour Collection book entry for a given marriage record. That page currently cites something silly like "uncited internet date." Do I delete that? What if it says something second hand like so-and-so's Worldconnect file? What if it cites the Ricker collection discussed above? (the latter being the real situation) I'm one of those 'highest and best' folks who only cites the best source I've got, unless there's a conflict, but I know thoughts differ.
• What are we doing for really broad sources? Some that come to mind are 1) periodicals like NEHGR (where the source cited should be an article, but I don't think we want pages for every article, despite the MySource result equivalent to that); 2) Ancestral File/WorldConnect/IGI/etc.; 3) the Barbour Collection, which currently has one page for all of Connecticut's VRs, which seems a little odd.

--Amelia.Gerlicher 19:03, 14 July 2007 (MDT)

Woo! Amelia, you're on a roll tonight! First question: sometimes it's not real clear which source is superior. I'm happy to just add improved sources without deleting the place I first found out about it.

But it's not just what I looked at or how you want to do your sources on people you upload. If this is a collaboration where we aspire to create the definitive page on Mr. X, don't we want to cite the best information available? I would prefer to see a minimum of the best possible cites, say a site to the parish register rather than having 18 cites to various articles and books because so much has been written about a particular guy. And I also would prefer not to see 10 utterly proper cites that say "parish register, viewed at X library on Y date", so the "what I looked at" question has some weaknesses here too. While I know there will be gray all over when you only have the articles and not the register, a theory on whether we're going for all possible or the best possible would be useful. But again not sure whether that's within the scope of what's going on here.--Amelia.Gerlicher 22:28, 14 July 2007 (MDT)

Second question: I would actually vote for citing individual NEHGR articles. In NEHGR, the unit of actual use is the article. To cite only NEHGR would be like citing "Donald Lines Jacobus" without being able to distinguish what it was being cited. Am I wacko here?

Worldconnect databases have titles and authors. PRF submissions have authors. Why can't they be cited that way? I'm not sure if there's a useful disambiguation for IGI, but my impression is that some submissions there are a lot more credible than others, e.g. the "extractions" are really transcriptions in my book.--Hh219 19:46, 14 July 2007 (MDT)

It seems to me that listing all WorldConnect and IGI submissions would be a fruitless exercise that would overwhelm the source search and make it unusable, while adding virtually nothing to the site. That does not mean that the databases can't be cited by author, which they should be.--Amelia.Gerlicher 22:28, 14 July 2007 (MDT)

Hmm, is it possible that we could think about this in terms of "usable units" for source pages? Obviously it's possible to have the source page be "New England Historical and Genealogical Register" and whenever you cite an article, you would specify the issue, date, article, and page number in each citation, not on the source page. My feeling is that that leaves very little work for the source page to do (what can we usefully say about NEHGR that would apply to 1860 and 2007 both?), and an awful lot of work for the individual citations to do. At the opposite extreme, it would surely be nuts to give each page of each issue of NEHGR a source page! To me, giving individual articles their own source pages (counting a multipart series as one article) is a natural middle-road choice, and they can be cited in the author-first style as well.

Looking at it this way, individual WorldConnect trees and Pedigree Resource File submissions (and freestanding family web sites for that matter) are analogous to NEHGR articles, just at a lower level of reliability. Plus they have authors and (in the case of WorldConnect) titles. Again, useful discussions of quality take place at that level. (I agree with Amelia there's not point in trying to do this with IGI or Ancestral File; if distinctions can be made other than "caveat lector," it may be easier to do so at the citation level rather than the source page level.)

I think a case could be made that the essentially unregulated and wildly variable contributions to WC and PRF don't merit such treatment. But I don't think "overwhelming the source search" is that argument, especially if the source search can be designed to exclude certain kinds of sources. (Dallan??) The argument would be that having individual source pages for particular WorldConnect trees is a waste of time because most of them aren't very good. The counter-argument would be that people in fact use them all the time and need to be educated, and if someone cites one as a source, then its own source page would be the ideal place to discuss why or why not that particular tree is reliable, whereas a source page that was for all of WorldConnect wouldn't. --Hh219 08:19, 15 July 2007 (MDT)

I think we could allow people to search sources by type of record, so they could omit certain types of records (say family trees).--Dallan 19:35, 19 July 2007 (MDT)

I can see your point on articles, and keeping the goals of the pages in mind, it does make sense to have pages on articles. I think (Dallan will have to tell us) that they would have to be created by hand rather than pulling them automatically from somewhere, however, which means we would only get pages on articles that someone has actually used, a good starting point. On WC, however, I still think it's a waste of resources (WeRelate's, ours, the end users) to include each file as a page. If I want to know what files in Worldconnect [etc.] have my surname, I'm much better served by using the search over there that will give logical results. I can't think of anything useful that a source page would add to that process. Even a space to comment that such and such data is wrong/sloppy is better done through a post-em. And the most likely workflow for an enduser would be to run a search for a surname, get a source page, and then have to go WorldConnect and search again to see if the right people are actually in the database.--Amelia.Gerlicher 16:25, 15 July 2007 (MDT)

Wow, what a great discussion! I started out thinking I had good opinions on several of the issues raised, but after reading the arguments on both sides I'm not so sure anymore. A couple of thoughts.

• It's going to be difficult to keep people from creating separate sources for the same record set if they are published under different titles and authors. I don't think we can preclude this, although we could redirect or at least link these source pages together.
• Another goal of the source index is eventually to help people know what sources are available when they get stuck -- kind of like a community-driven GenSmarts. In order to achieve this, the sources need to list the time periods, places, surnames (but listing the major surnames covered is important only for things like family histories, biographies, etc.), and record types covered. I think this means that having separate sources for each NEHGR article would be required. We might be able to get a list of the articles from NEHGS, but the coverage information would still have to be manually entered.--Dallan 19:35, 19 July 2007 (MDT)

OK, I'm convinced! Anybody else have a problem with WC, PRF, IGI, and AF each having a single source page, and those of us who cite them (blush) make any needed distinctions in the actual citation rather than on a source page? While we're at it, Amelia, what's your thinking on family trees that are on line but aren't part of the excellent search functions at familysearch.org or rootsweb? --Hh219 16:50, 15 July 2007 (MDT)

I'm a little more divided. Those trees lack the accessibility of a specialized search, but they still have both Google and the WeRelate web search to find specifics. They have no post-em feature, but a source page would be kind of an awkward place to complain about the few errors one might identify. For descendant studies, there might be some usefulness in a Source page, because it could be made findable and identifiable. But "Susie's Kin" that has four generations of random people isn't going to be useful. Perhaps we could strike a more limited ground where we discourage such pages unless the title of the website sufficiently identifies its contents, and the contents are all related to a particular surname, area or ancestor.--Amelia.Gerlicher 21:43, 16 July 2007 (MDT)

I'm in favor of the "mega-sites" like WC, AF, PRF having single Source pages. Regarding individual websites, there are currently around 100,000 websites containing family trees. Even though I've created source pages for many of them, I'm not sure that having the source pages is worthwhile. Having said this, I know a number of people have asked about creating source pages for their websites on WeRelate to make sure that Google can find them. Since Google crawls WeRelate, I've said ok to this point. Alternatively, these people could just as easily submit their URLs directly to google, but there may be some benefit to having an open-content list of genealogy websites that anyone can use. I could go either way on this -- denying them altogether, removing all of the automatically-generated source pages for websites that are currently in the source index but allowing people to manually add source pages for the websites that they want to, or keeping the automatically-generated source pages as well.--Dallan 19:35, 19 July 2007 (MDT)

Evidence Explained - Citing History Sources from Artifacts to Cyberspace [6 August 2007]

"That woman author" Jillaine referred to above is Elizabeth Shown Mills, and the new and greatly expanded edition of her book Evidence Explained - Citing History Sources from Artifacts to Cyberspace is going to be released at the FGS Conference in Fort Wayne next month. I believe that this discussion on source citation format and page title format would be more valuable after we have had the opportunity to view her book. Mary--Kittydoc 08:17, 14 July 2007 (MDT)

I've been studying the first edition of "Evidence!". One issue is that it primarily focuses on how to cite individual sources (e.g., an individual birth record), although the bibliographic entry format that she gives is often pretty close to what we want I think. I've tried to follow the bibliographic entry format in the WeRelate talk:Source Committee/Examples page. I agree it will be good to see what the second edition has to say. I'll pick one up at FGS next month.--Dallan 19:35, 19 July 2007 (MDT)

Just FYI... rumor has it that the new version of Legacy, due out in October, will have source templates that follow the guidelines of "Evidence!". --Ronni 23:51, 5 August 2007 (MDT)

Sources in uploaded GEDCOM's

Um, I don't see anyone talking about what to do about sources that are included in the uploaded GEDCOMs. How will those be reconciled with these sources you're describing here? Jillaine 19:38, 19 July 2007 (MDT)

The plan is to continue to create personal MySource pages for sources included in uploaded GEDCOM's, unless the title of the source in the GEDCOM starts with "Source:", in which case we'll assume that it refers to a community Source page. At some point I'll see if it's feasible to automatically match GEDCOM sources with community Sources. But it will require pretty well-described sources for the match to be reliable. I'm open to suggestions on how this could be better.--Dallan 21:37, 19 July 2007 (MDT)

Side comment: Oh. Damn. I suppose this was sort of obvious, but what this means to me -- with endless hours already sunk in a database of 35,000+ people and 3500+ sources, is that uploading my GEDCOM to WeRelate won't do that much good, because every single source and every single citation will have to be done over if it is to be findable. By hand. Maybe if I were finished with researching, or if I were just starting, I would feel differently, but this makes me feel like the wiki promise is very far out there indeed. Nobody's fault, it's just hard.

--Hh219 06:14, 22 July 2007 (MDT)

I may be missing something here, so please tell me if I am. I don't see much benefit in changing existing GEDCOM sources to point to the community sources. Let's suppose that I have a source in my GEDCOM that gets converted to a MySource on upload, and that the source is really the same as a community Source. If I take the time to change the MySource to a community Source, then I get the following advantages (any others?):

• I can see usage notes that others may have added to the community Source,
• I can see the correct author, title, and other publication and location information that people have added to the community Source, but then I probably had some of this information already in the MySource,
• Others can click on the "What links here" link at the bottom of the community Source page and see that my Person page references the community Source,

Are these benefits worth the trouble for GEDCOM sources? I can see a lot more benefit of community Sources for new research. Rather than having to enter all of the information about the source so that others can determine how much to trust the source or look up the source for themselves, I just enter the title of the Source on my Person page and anyone can see the publication and location information on the Source page. Also, I can search community Sources and MySources at the same time to see what sources might be available to me when I get stuck in my research.

If people think that the benefits are worth it, we could try to automatically (or semi-automatically) match sources in uploaded GEDCOM's to community Sources sometime next year. I'm just trying to understand the benefits more clearly.--Dallan 20:45, 22 July 2007 (MDT)

I think it would be great if you could match some sources uploaded in gedcoms to the actual source pages, like you do with places. Realizing that I didn't like the treatment of sources is one of the reasons I've opted to add my stuff in by hand for the most part. (The other is realizing the duplicate problem in people and being too impatient to wait for the merge function.) In addition to the benefits you list, it will help usability by avoiding the creation of some duplicate MySource pages (which clutter search) and it should just look better on the page. I'd like to see the function identify and link to source pages from the title/author I upload, and keep the rest of my citation appended as detail on the person/family page. You'd have to be a little fuzzy on the logic, particularly for things like middle names and subtitles. One dream would be if you could pick out unique info like "FHL Film no." or a website and link accordingly. Two concerns: if you can't match the source, don't make a Source page (that would create a lot of junk, much in need of merging); and I'm more excited about merging people and overhauling search than this (which is where it sounds like it is on your list, so that's good).--Amelia.Gerlicher 20:55, 23 July 2007 (MDT)

Examples

Dallan has some examples on the next page.

The examples are very good and I have just a few comments about them:

1. The length of the some of the titles is going to cause troubles because of the number of characters we're limited to, isn't it? Especially when we start naming the "responsible agency" or URLs for websites in the titles.
2. I personally prefer using a period (.) instead of a comma (,) between author's name and title. It quickly gives a visual clue where the author's name stops and the title begins.
3. I'm not sure I like adding the "responsible agency" to the title or not. Would have to hear more pros and cons about that.
4. It seems like a lot of rules. Is there going to be some kind of "wizard" to help us in creating the title for the source?

--Ronni 11:37, 19 July 2007 (MDT)

You beat me to the punch on this :-). The WeRelate talk:Source Committee/Examples page should be treated very much as a "work in progress". Please feel free to modify it as much as you want. If someone wanted to enter additional examples from "Evidence!" that might be helpful as well.

• I think the rules for creating source page titles can be summed up as: the Source page title in the source author followed by the source title.
  • If there is no title, enter the title as a description of the item (e.g., "Journal" or "Minutes" or "Birth Registers").
    • When using a description of the item as the title, add years or other information necessary to distinguish the item from similar items after the record description (e.g., "Federal census, 1870, population schedule").
  • If there is no author, but there is a "responsible agency" for the records, enter the responsible agency as the author.
    • When using an agency as the author (e.g., "Bureau of the Census", or "North Carolina Division of Health Services--Vital Records Branch"), prepend the place covered by the agency (highest-level jurisdiction to lowest-level jurisdiction) to the front of the agency name (e.g., "United States, Bureau of the Census", or "United States, North Carolina, Division of Health Services--Vital Records Branch").
  • If there is no author or responsible agency, it is ok to omit the author. The Source page title is then just the source title.
  • If we want to have different titles for different editions, different transcriptions, or different translations of the source, we could add the edition/transcription/translator to the end of the Source page title to make it unique (e.g., "United States, Bureau of the Census, Federal census, 1870, population schedule. Ancestry.com").
• The rules for filling in the publication information are more complicated. This is where I think having different templates for different types of records, where each template would have different fields especially for publication-related information, would be the most help.
• We need to keep Source page titles to 150-200 characters. The MediaWiki software has problems with page titles longer than this.

I'm open to suggestions for improving the Source page title rules. The above is just a straw-man.

Regarding adding the name of the "responsible agency" to the Source page title, instead of just the place covered by the record set, I changed this in the examples because it makes the Source page titles follow the first part of the bibliographic entry format in "Evidence!". It also makes the Source page title format the same for an item authored by an agency and an un-authored record set with a responsible agency. For example, are the minutes of the Flint River Baptist Association an authored item, or an un-authored record set where the Flint River Baptist Association is the responsible agency? If we include the responsible agency as part of the Source page title, then it doesn't matter because the title is the same. But it does make the titles longer, and it introduces additional possibility for error when entering the title.--Dallan 19:35, 19 July 2007 (MDT)

Some comments on the examples:

• Generally, I agree with the comment up above that some of these titles end up way too long, and I think worrying about the "responsible agency" is most of the cause. I do not think "responsible agency" belongs in the title -

1) it's too long;

2) it doesn't help identify the source in most cases (adding "county clerk" for birth registrations that only exist in one place, for example)

3) It's going to harm usability when looking at the dropdown when typing in a source on a person page or at a list of sources for a locality (compare "States, Texas, Archives Division, State Library. Confederate Pension Records, Index, tsl.state.tx.us" vs. "United States, Texas, Confederate Pension Records, Index")

4) It's too much like a respository and thus 5) it's going to be hard to figure out/format correctly.

Admittedly, that's all based on my opinion that a title should convey the minimum useful information to reasonably identify the source, and that additional information like publication/whereabouts goes on the page itself or in the citation where it's used. And based on the ideas discussed above that we're putting multi-volume/edition books on the same page. I recognize that the generaly rules we formulate for automating make some things ugly in the service of the greater good. But I do think that the purpose of these page titles is not to conform to proper citation formats. It should be to identify the record. For non-authored works, location is far more important than responsible agency.

• Also generally, that example list is way, way too long for general use. We're going to need a 'top three' list of examples that can go on the help link, with further information elsewhere. There are also a number of items on the list that appear to make a distinction without a difference.
• Specific comments:

• Census I thought we didn't want separate pages for online images of the census vs. the films. And why is there just one source for all of 1840? We're not doing county-specific? (I'm all in favor of reducing clutter, but I don't find the idea of one page for the 1840 census useful at all. I'd much rather know where I can find copies of the Jasper County, Missouri 1840 census, for example).
• National Archives Should not be in the title because it's ultimately a repository. And the title for Bounty-Land File (federal, unfilmed) is totally out of control. Pension and Bounty-Land Warrant Application Files, 1800-1960 would be sufficient. The rest can go in the fields. Of course, this is going to be such a rarely used source that I won't be heartbroken if it's unusable.
• Personal websites I think the source title should have the website homepage title and organization/person author if available. No url, because they can get extremely lengthy. AND there should be a rule that if the website has no title, it doesn't get added as a source page. What would be the point? Who would find it? How would it get used? If it has no title because it covers too much material to encapsulate in a title, then it's not going to be useful to have as a source page anyway, and can comfortably stay a MySource if needed.
• Bibles Who's the author of a bible? Do you mean owner? Original owner? Current owner? What if you just know it's old and has Adams family information in it? --Amelia.Gerlicher 11:16, 20 July 2007 (MDT)

I second Amelia's motions (though I'm conflicted on using responsible agency -- the nice idea is not to have to think about whether it's "authored" or not). Especially the one about censuses. Here's another reason if we need one. For whatever reason, a lot of folks will just cite the source that's given, and not put in any citation details when they cite it. If the standard source is "United States, 1850 Census" that's how they'll cite their info for Uncle John. If the standard source is instead "United States, Illinois, Putnam County, 1850 Census" then the rest of us will at least have an idea of where to start digging if we need to find the entry for Uncle John. (I know, you can take this argument to the extreme by making every individual household a separate source, but then surely the search function would collapse with a massive coronary.)

BTW, can anyone make a rational defense of treating census citations differently before and after 1840?

--Hh219 06:14, 22 July 2007 (MDT)

More specific example comments:

• It takes a long time to go through these but it's wonderful to have them to work on. Thanks, Dallan.

• Yes, it's a lot, but hopefully it will be intuitive which one to pick, and having a detailed template will certainly encourage me (for one) to actually try to make sure all the relevant information is provided! And to realize what information is relevant!

• I wish the author name could be in the same format in the Source Page Title as it is in the Source Field for author. I mean, it would help not have to remember, "last name first in the title and last name last in the source field."

• All these templates include the complex "location" box allowing for multiple locations, as previously discussed, right? Hm, well, maybe not, like the articles. On the other hand, in the case of Bible records, often copies or transcriptions of the record are in repositories other than the original Bible, so we'd want a location field set there, in addition to the unique fields about who owned the original Bible and when.

• Is there any reason whatsoever to distinguish between local and state birth registrations? Aren't the templates going to be identical?

• Even when it happens not to be appropriate for the given examples, surely we want pretty much every source template to continue to include fields for surnames covered, places covered, and year range covered? This is an example of where a wiki source list can serve more functions than just making it possible to produce a useful citation -- it's also for searching and finding relevant sources themselves.

--Hh219 06:45, 22 July 2007 (MDT)

Summary? [8 September 2007]

At some point (say September) would be helpful if someone could summarize the questions that have been raised so that after discussion we could "vote" on the issues.

I'll ask the people at ACPL to weigh in on these issues once the FGS conference is over.--Dallan 21:37, 19 July 2007 (MDT)

Alright, it's September. I'm going to take a crack at summarizing. It occurs to me reading through this that there are several types of things going on, so I'm going to go for a few subtopics. I'm going to sign up here and invite everyone else to edit this list below. If you want to explain why, then comment up here.

These are based partly on the table of samples, partly on the discussion above, and partly on the directive from Dallan that you should be able to start typing and get a list of sources. That means starting with something obvious (author or place - so I came up with some new things not otherwise discussed). And we want to make this as simple for people as possible. --Amelia.Gerlicher 13:36, 8 September 2007 (EDT)

General Rules for Source Page Naming

(i.e. How to name a new source)

- If it has an author: Last, First. Title: Subtitle

- This includes: Books, Articles (title is article title), compilations that list a single person or entity as the compiler or editor, and webpages when entered.

- Two authors, use both names (Last, First & First Last). More than that: Last, First et al.

- No Author: general rule is either just Title or Place. Title

- Vital records: Country, State, Locality. Title of records

- The title is either the title of the book or microfilm (like the ones in the FHL library) or a common sense title ("Longview Cemetery", "Birth Certificates")

- If the title or record set includes only certain years, that goes in the title (i.e. 1880-1960)

- Census records: Census. Country, State, County, year

would this be "Country, State, County. Census, 1850"?--Dallan 17:10, 9 September 2007 (EDT)

- One page for each county in each year? (We could have one for each year or one for each year for each state or one for each county across all years, but the reality is that people cite to county and year)

- No differentiation for transcriptions vs. images (because 1) it would be very useful to have all transcriptions listed in one place and 2) most people don't distinguish where they found the record the first time)

- Unauthored book or other source: Title

- Might include city directories, sources like WorldConnect and IGI, and CD-ROM compilations of multiple state records.

- No "responsible agency" because it's long, confusing (for trying to correctly name a page), and much like a repository.

Special cases

- Books with long subtitles: Cut off the subtitle at the most logical place within the first ~10 words. For example, something like "The Adams Family: The history of Henry Adams and his descendants in Massachusetts; his son John Adams and his descendants and the Adams family in Rhode Island" gets truncated to "The Adams Famiy: The history of Henry Adams and his descendants in Massachusetts". Perhaps the system could truncate at the first semicolon? (And can we get the system to 1) use title case and 2) omit the extra space before the colon for subtitles?)

Yes, I think we can get the system to set the title of a created page to be in title case and to omit the extra space before colon.--Dallan 17:10, 9 September 2007 (EDT)

- Articles. Each article in a genealogy periodical gets its own page. (Last, First. Title). The template obviously needs to include the periodical title, date, and page numbers. There should be an exception for all those articles in the early NEHGR that are just vital records transcriptions; they should be entered as vital records.

- Second editions of books. I'm going to think of this as if I'm looking at the book --

- Same title and author = same page. So a second printing (or later reprint) of a book doesn't generate a new page.

- New author "revised by" or "edited by" = new page using new author and new title. Our collective human power will have to link the two.

- Multi-volume books: Based on title and author as much as possible.

- Same page if they all have the same author.

- Same page if they all have no author and same title.

- Different pages if they have different authors.

- Same page if the titles are different only by year or part of alphabet. Title omits years/letters (i.e. the Great Migration books are all on one page instead of separating 1630 A-B from C-D)

- Different page if volumes cover different localities (i.e. the Barbour Collection are all different books)

- Vital records: The source page is for a set of vital records for some particular locality that exists as a set in an accessible form: A locality's file of death records, a book of vital records published for a town, a cemetery. All versions of that set are on the same page:

- the original book,

- images of that book on Ancestry or Google Books, etc.

- a transcription of those records on a webpage,

- the filming of those records for the FHL,

- the transcription of those records in the IGI.

The reason for this is to collect information about where to find a source (the vital records for X place) in one spot, which helps further research and allows comments about the virtues of various transcriptions. And it makes it easiest to enter the source for people entering sources (or for the system trying to match them)

This rule does permit different pages for different original "sets". For example, the Barbour collection book for a particular town is a different source than the original records for that town, because Barbour drew his info from a variety of different places. A CD that compiles all known information about X town gets its own page, even if there are pages for each of the record set it used. But it two sources got all of their information from the same place, they should be on the same page.

- Compilations of user-submitted records (Ancestral File, IGI, GenCircles, WorldConnect, etc.): one page for each. The detail to the source should be more specific where applicable, but there should not be a separate page for each batch/file.

- Personal websites: Author. Title or just Title. But wherever we put this list, we add a disclaimer "Please do not create source pages for websites with generic titles ("Kim's Kin" "Ancestors of John Doe"). If the title does not convey anything about the content (or if there is no general title), consider using a MySource page instead."

We have a lot of websites (around 50,000) currently in the source wiki that don't follow this format, but I'm not against removing them and letting people enter website sources with a proper title when they need to.--Dallan 17:10, 9 September 2007 (EDT)

- Gedcoms. Often cited by people, but should not get their own pages. Same with "Research of Jane Doe" for when somebody [e]mails group sheets, etc.

Templates

There's really not much discussion of this, but a few thoughts:

• Keep the surnames and times covered.
• The existing template works pretty well for books and similar sources.
• Given the rules about vital records above, that probably needs its own template with information about the original source and other locations.
• Articles will need their own template.
• Repository and URL should have multiple lines

Issues related to search [5 September 2007]

The source "form" includes fields for surnames and locations, but what about a keywords field? That could be used to search for sources related to nationalities or ethnic groups or other general topics.

--Moultriecreek 16:28, 3 September 2007 (EDT)

There is currently a keywords field for searching sources. I'll make sure that we don't remove it.--Dallan 10:00, 5 September 2007 (EDT)

What's an author? [18 September 2007]

Is a compiler or transcriber an author?

Is an editor?

Is an agency that compiles records an author? What if it's "Town Clerk?

Is the person who published a set of vital records an author? Or would we rather have vital records named using the place? (The advantage of the former is specific identification, the advantage of the latter is easier finding in categories and easier combination of subsequent reprints (where the original extractor goes missing)

I could vote either yes, yes, no, no, or just no entirely for consistency, but I don't feel strongly either way. I'm a little worried about agency/association names getting out of control or unhelpfully generic. And I like the idea of having vital records titles identify the place--Amelia.Gerlicher 14:04, 15 September 2007 (EDT)

My vote would also be yes, yes, no, no, for the same reasons as Amelia.--Dallan 11:36, 18 September 2007 (EDT)

Prioritize the basic sources? [20 September 2007]

Pardon my interloping on a conversation that's been going on for a while now (I'll do my best to catch up), but I was frustrated with the current state of the Sources, and was about to start a thread on the Watercooler, when I noticed this Committee page. What I'd like to suggest is that whatever plans are afoot for getting the Source namespace organized, it may be a good idea to focus initial efforts on the most-used basic sources, chiefly the US/UK/Canada decennial censuses, and a few other workhorses like the SSDI. I say this from the pragmatic surmise that a small number of sources will account for a large number of source citations, and the observation that the population of our WeRelate user community is swelling. I read above that Dallan has a plan in mind for an automated conversion of the existing Source pages to a new format once that format is agreed upon, but I think that the basic sources require some manual intervention before anything automatic could have a good result. For example, the existing source pages related to the US Census are a haphazard crop, and don't seem to follow any organized convention. Some target particular years. Others target particular websites. Still others target particular counties or regions of various scale. A newbie coming to this site and wanting to cite a particular census would be hard-pressed to find what he needs by searching the Source namespace as it stands.

I was tempted to follow the Wikipedia motto to "be bold", and go ahead and create the Source pages that I thought ought to be there, but on discovering this conversation, I see that what I had in mind for the census was different than what others here are suggesting. I had been thinking of one Source page per country/year, e.g., "US Census 1880", "UK Census 1841", etc. This has the advantage of requiring only about two dozen Source pages to be created to represent the US, UK, and Canada censuses. However, it sounds like the going suggestion here is to have a Source page per county/year, e.g., "Census, US, Wisconsin, Lincoln County, 1880". That requires several orders of magnitude more Source pages (i.e., thousands vs. a couple dozen), and I wonder to what advantage. I think it was hh219 above who said he (she?) didn't see the point in one census page for an entire country/year, whereas s/he would like to see the county-specific census page to provide info on where that census can be found. My thinking on this was just the opposite. I imagined a very useful "US Census 1880" page. For example, an explanation of the census form, and the various annotations that can go in each column. That information is specific to the year, but common across the nation. Conversely, while I understand that local portions of the census may be available in particular local repositories, entire national censuses are widely available from a number of repositories (both physical and online). I'm thinking that a county-specific census page would largely duplicate common information, and would end up saying things like "the entire 1880 census is available in these places ... and in addition, the Lincoln County portion of the census may be viewed at the Merrill Branch of the Lincoln County Library."

I share hh219's eagerness for local knowledge -- what are the repositories of interest in Lincoln County? what are the useful pieces of information unique to history/genealogy in Lincoln County? I guess my thinking is that the Place pages are a good place to put all that, in the scheme of the overall wiki. Place:Lincoln,_Wisconsin,_United_States could tell you where in Lincoln County to find local portions of the census, as well as whether some sections of it are faded, or that the enumerator who did the town of Tomahawk had some particular quirk in how he annotated those census sheets. It could tell you where to find the microfilms in the Merrill Library, and that the pastor of the Brandon Lutheran Church took compulsively copious notes on the congregant population from 1889-1913 which can be found in a university library in the next county. I urge everyone to adopt a county Place page. :-) (And if you've any interest in Scotland, please see here!)--TomChatt 03:39, 19 September 2007 (EDT)

Tom, I though you almost had me almost convinced...and then I thought about it more and decided that you do have me convinced. I do think that there are significant resources out there that are not the "full 1880 census," and, particularly among the free resources, most are just one county or state, and we need a place to collect those. But we could do it on the county page, which would have the advantage of being just one page for the county instead of 10 to draw users and editors. One thing that may help reduce the downside (only having a link to the 1900 census in the source cite, which won't be very specific to that citation) is to do an example or part of the template that adds a link to the county page to the source cite. Otherwise, if the person types in a place in an event field, it's likely to link to the town instead of the county where most of the records are likely to be.--Amelia.Gerlicher 11:07, 19 September 2007 (EDT)

I'm a big proponent of using Place pages to capture local information. What really turned me on to their potential was discovering what GENUKI has done with their place pages for the UK. I find them invaluable and refer to them again and again, but I can't add or make corrections to them directly as I could on a wiki like here. To get an inspiration of what WeRelate's Place pages could evolve into, take a look at this page for Aberdeenshire. Now imagine the WeRelate Place page for your favorite county containing that sort of info. Let's make it so! --TomChatt 02:36, 20 September 2007 (EDT)

Regarding Census Sources - need for both general and local [2 June 2008]

I think we need both the general (ex. 1910 US Census) pages as well as the local state/county/year (ex. Illinois, Franklin, 1910) pages. We need the county pages because some county census records have been transcribed and annotated like this one for Franklin, 1910. This version only covers a township in Franklin County, but this is the kind of local information I think we are looking for.

Once the Source pages are standardized and duplicates combined, we can truly realize the benefit of the "What Links Here" function at the bottom of the pages. This is evident when you do this now (try it!) on a Place page. The only problem is that the results are not ordered by type, but by when they were created. This needs a dropdown box so that we can see all the Person, Source, Place, Family, etc pages. If you haven't been using the "What Links Here" you will be pleasantly suprised.

I can certainly see both points of view on this. But I'm trying to think of what makes the most sense from a "wiki taxonomy" point of view (which has both information science and technical/programming aspects to it) as well as a pragmatic manageability point of view (how do we get it done?). I agree with the spirit of "let's do it both ways", but I don't think we can do it both ways both in the Source namespace, otherwise one would have to make a choice each time you added a source whether to select "US Census 1880" or "Census, US, Wisconsin, Lincoln, 1880" for the "source" field on the page. However, I think there is a good wiki solution to accommodate both, either way we decide to organize the census Source pages. We have:

• Plan A: census Source pages at national level, local information about the census is put into the county (or state or town) Place pages as appropriate
• Plan B: census Source pages at county level, common information about the national census can be put into a Category page (e.g., "Category:US Census 1880") that all of the Source pages belong to.

I think that either plan could be made to answer most if not all of what we want. The difference I worry about is that Plan A can be implemented by creating about two dozen Source pages (one for each census year for US, UK, and Canada). One or a few of us could do a high quality job of that in a reasonably short time. Plan B requires many thousands of pages to be created (one for each county per year). If we wanted it to be accurate and complete, it would require a bit of automation to do, and it would require having the appropriate information as input, namely a good list of states/counties per census year. (We couldn't just use the existing Place database because it wouldn't have the time component: states and counties evolved over time, so the states and counties in the 1830 census aren't the same as those in the 1930 census.) Perhaps that's not as hard as I think, Dallan would have to answer that. (It's certainly doable, as, e.g., ancestry.com has done it, though whether via an automated process or with human editorial hard work, I don't know.)

Things that are a wash with either plan:

• figuring out what to do with the existing haphazard bunch of census-related Source pages
• human editorial effort required to actually put in the useful local information is the same, whether it goes on a county Place page or a census-county-specific Source page
• ditto for the national census info, whether it goes on a national census Source page or a Category page

I know that concern has been expressed about making sure that source citations are in the proper format. I believe that if the information is structured sufficiently, the proper format could be generated with either Plan A or Plan B. Plan A, however, would probably entail adding a "Place" field to the "add source" template on Person and Family pages. (Or better still, somebody had suggested a "wizard". That would be a great way to help people consistently enter appropriate info. I'd say that adding a census cite looks like a top candidate for a wizard!) With the right structure, the wiki could automatically put the right information pieces in the proper format when presenting a Person/Family webpage, or when exporting to GEDCOM.

Another consideration is the ease of entry when adding a source. When we had a similar discussion amongst WeRelaters active in Scotland about the best way to name the Source pages for the parish registers (one of the fundamentals of Scottish genealogy), some expressed concern about making it easy to enter, taking best advantage of the drop-down menu completion feature. If we had only to enter "Census, US, 1880" in the source field, and then the county name in a separate "place" field, each of which had appropriate auto-completion, that would make it easier on the fingers. ---TomChatt 02:33, 20 September 2007 (EDT)

Good morning everyone,

I have not formed any opinion on how census sources should be formatted, but I have some observations.

• Census Source Level: First and foremost I need to locate the source almost immediately. If one chooses to only have the census source for country and state, then there will be many more links per source page. The user will also be required to type more data into every citation to show the county.

• Type of census record: Should one include all types in one source? Ancestry, Rootsweb, Heritage Quest, FHL microfilm, county level, etc? This again requires the user to type more data in the citation.

I'm somewhat of a heretic in this regard, but I don't think that it really matters whether the census image comes from Ancestry, HeritageQuest, FamilySearch, etc. As long as you have the state, county, and page number, you can find the image. Note that I'm talking about images. For something like Rootsweb (thinking about the numerous transcriptions on there), the source is the Rootsweb site, not the census. Of course if it is an actual image, then the census is what needs to be cited. --Ajcrow 18:51, 25 January 2008 (EST)

Each person has their own method. I do put where I accessed the image; but I can add this at the citation level. The source format needs to be flexible to allow for different methods by users. --Beth 08:48, 26 January 2008 (EST)

Remember that the "source format" we're talking about is primarily the source title -- the title for the page that will have information about that source. I think it would be unnecessary, excessively technical, and destroy much of the wiki benefit to have separate source pages for Ancestry, HQ, etc. census pages. Once we determine a level for which we have census pages, each page (just like all other sources) can easily list the available repositories, and anyone can comment on the differences between them there. If someone wants to add that information to the citation detail, they can, but I think even that's unnecessary, since there's very rarely going to be a difference.--Amelia.Gerlicher 21:05, 26 January 2008 (EST)

• Also when formatting the sources for census keep in mind that in the later census years, there may be multiple roll numbers per county.

I'm confused as to what impact you're concerned about here.--Amelia.Gerlicher 21:05, 26 January 2008 (EST)

It may not, but I envision having the census data sourced at the county level and at some point having every person that has a source citation for that census year and county being automatically linked to that source page. That could be most helpful in a study of migration patterns of families. I assumed that on the source page one would site the National Archives microfilm number and roll number or numbers. Then one could specify the roll number in their citation. --Beth 09:23, 27 January 2008 (EST)

• Uploaded gedcoms will probably have census data entered in at least 20 different variations.

--Beth 07:44, 25 January 2008 (EST)

I have created a sample census source for you to examine and agree or not. Source:United States. Texas. Comanche. 1900 U.S. census, population schedule. This is the citation: United States. Texas. Comanche. 1900 U.S. census, population schedule, page 249 B (stamped), fair quality, 18 June 1900, De Leon, NARA microfilm publication T623, roll 1623, SD 4, ED 34, dwelling #121, family #123, Alfred C. Coker; digital image. Ancestry.com. (www.ancestry.com) --Beth 20:32, 28 January 2008 (EST)

Can we agree on consensus that

1. The format for a general census page is United States. 1XX0 Census Population Schedule
2. The format for a county census page for those that wish to have them is United States, State, County. 1XX0 Census Population Schedule
3. Both types of pages are independent of the medium or location of a transcription, with Ancestry, HQ, FamilySearch, etc. listed as repositories.
4. Citation using either adds sufficient detail to locate record (i.e. county, locality, page, ED, etc.). (What that detail level is is TBD and unnecessary to this specification.)

I would suggest that we also set up categories for each census year that are used to link the general pages to any county pages or pages for other schedules; it's easier than adding and organizing what could be hundreds of links from the general page to the county or state pages.--Amelia 00:43, 25 May 2008 (EDT)

That looks like a good format to me. I really like the category idea. (Although overall I think more explanation/examples of what categories are and how they function needs to be given. (Geez, that was a bad sentence!) One point in the above suggestion that I would change is in the third point. A transcription of a census record is NOT the census record. If I've used The 1820 Census of Fairfield County with Indexes by Rebecca Huddle (2005, Fairfield County Chapter, OGS), which is a transcription of the 1820 census, my source is the book, not the census. --Ajcrow 08:55, 25 May 2008 (EDT)

Technically, you're right, but I don't think we should get all worked up over whether someone uses that source page or the 1920 source page in the citation, since they're going to be identical 99% of the time (hopefully) and the 1820 census itself will be far easier to find and thus a more useful citation. And all those random internet transcriptions of counties should be listed as links on the county census pages, and not as their own source pages.--Amelia 11:32, 25 May 2008 (EDT)

People are going to cite what they're going to cite regardless. However, if we are ever going to write some "how-to"s for sourcing, I think we should be clear. --Ajcrow 11:50, 25 May 2008 (EDT)

Amelia your format is okay with me, but I thought we already had a consensus on the format that I have used recently. What matters is that I wish WeRelate would come to a final and definite consensus so I don't have to continue to rename my census source pages on the county level. --Beth 09:38, 25 May 2008 (EDT)

I didn't mean to change existing consensus; I thought the result of your example and other discussions was that periods were out and title capitalization is in. The inclusion of U.S. or not was just something I dropped as redundant, but if it will save effort, it can go back in.--Amelia 11:32, 25 May 2008 (EDT)

Amelia, this was the last final consensus <g> as I understood it.

United States. XXXX U.S. census, population schedule

United States, state name, county name. XXXX U.S. census, population schedule

But actually I have to agree with you that it would be easier to remember the format if we capitalized everything and removed the comma between census and population schedule. Regarding the usage of U.S., maybe we should leave the U.S. in the format. If we do not how are we to distinguish between a federal and state census? And just maybe, hopefully Dallan can rename my census pages. --Beth 12:21, 25 May 2008 (EDT)

You're right on the state/federal census issue - I had been thinking it was unnecessary at the United States level, but I didn't think about the county level. Which brings us to this format:

United States. XXXX U.S. Census Population Schedule

United States, state name, county name. XXXX U.S. Census Population Schedule

(And even if Dallan can't rename, your format Beth is close enough that the alphabetization will be nearly identical for search/drop-down purposes.) --Amelia 14:59, 25 May 2008 (EDT)

Sounds good to me Amelia; now if we can get everyone else to approve the final final census format; maybe we should add this to one of the tutorials on sources.--Beth 15:23, 25 May 2008 (EDT)

This format looks great to me. We need to publish it as an "official" recommendation. Do you think one or both of you could add it to WeRelate:Source page titles? I'd really appreciate it.--Dallan 01:44, 1 June 2008 (EDT)

Dallan, I will add this as an official recommendation. I think Amelia is having her baby. --Beth 02:16, 1 June 2008 (EDT)

I'm so glad to hear that's what's going on! I saw her comment on the watercooler that she'd be away for an indefinite period and I got worried.--Dallan 15:29, 2 June 2008 (EDT)

Trying an actual example [28 January 2008]

I thought I would try adding my first actual Source page, and see if I'm doing it as the community expects. I've been working with a photo-reprint of an old book called Memorial of John Slafter: with a genealogical account of his descendants, including eight generations, by the Rev. Edmund F. Slafter, A.M. Turns out there were already two Source pages for this book, one originating from the FHL microfilm of it, and another originating from online images of the book available at ancestry.com. If I understand the consensus correctly, the feeling is that there ought to be a single page for the book, which identifies various forms it is available in (FHL film, online images, photo-repro, etc). Noting that both the existing pages were "stubs" that had nothing linking to them, I renamed one of them to conform to the format we've been describing here, and I changed the other one to also redirect to the same place. I then fleshed out the page a bit. Does that sound like the right thing to do? The new page is here and the original pages are here and here.

Some observations:

• I thought it would be better to drop the lengthy subtitle for the purposes of the page name. It seems like there ought to be a field on the edit form for the actual Title. I ended up just putting it as part of the free-text description.
• Is the author field to be filled out bibliography style ("Slafter, E.M.") or in straightforward form ("E.M. Slafter")? I went with the latter.
• Publisher field should be bibliography style ("Boston: Dutton and Son, 1869.")?
• URL field - it would be nice if you could enter more than one, as is relevant in this case - or should that field be left blank except for Sources which are websites
• Call number?? Mightn't that vary by library? Or are those things universally standard?

---TomChatt 04:56, 28 September 2007 (EDT)

Thank you TC! As the source page is now pre-formatted, this looks like an admirable job to me. I'm thinking the reformatting will make it easier to add multiple URLs and also repositories. Other questions (like the way we do the author's name) just have to be decided. --Hh219 17:32, 28 September 2007 (EDT)

on Call numbers, it is my understanding that these are universal, being assigned by the Library of Congress. But I am not a librarian so our members from Fort Wayne Allen County Library (or another library) can better answer that. I did find this online "Library of Congress call numbers. (This and the Dewey Decimal system are the most common call number schemes used. Dewey is more commonly used in public libraries; Library of Congress is more commonly used in larger research libraries in the US, and a similar system is also used in Canada.) Library of Congress "call numbers" actually start with letters, each of which represents a different topical area" found at: The online books page --Msscarlet1957 07:58, 15 October 2007 (EDT)

Some observations;

1. Title of article. Presumably one of the uses of "Source" articles is to reference where information comes from when it is given in a person article, (etc). For that purpose a very simple title is much better than a more descriptive lengthy title---limiting the amount of information needed to accurately identify a work is a good thing. A title like "Emma Siggins White, 1902. "Descendants of John Walker of Wigton Scotland, together with....", is cumbersome and unwieldly--especially if it has to be inserted in a number of places. A title like "White, 1902" is a lot easier to use. This "author, date" style also conforms to that commonly used in many fields for referencing specific sources.

2. Tom's Article. I thought this was a good, helpful layout. I'm using the style he suggested for my own work. The format would benefit from the addition of a biliographic entry. That would give the following standard elements:

• Authors Name---e.g, White, Emma Siggins
• Date of Publication---e.g., 1902
• Title---e.g., "Descendants of John Walker of Wigton Scotland, together with....", in quotes or italics
• Place of Publication---e.g., Kansas City, MO
• Publisher---e.g., Tiernan-Dart Printing Company
• Pagination---e.g, xxx+732pp

There is probably a need for different patterns for different types of documents---e.g, books, journal articles, newspaper articles, diaries, etc.

3. Input boxes. When a page is being created you get some of the following input boxes:

• Year range: -
• URL:
• Author:
• Publication info:
• Call number:
• Type:
• Repository name:
• Repository addr:

I'm not sure what is meant by some of these. (There's probably a need for those "blue question-marks" that explain what is meant.)

• Some, like "Author" are clear enough, but is the convention Last Name, First Name Middle Name? or is it First Name Middle Name Last Name? etc.
• Is Call number "Dewey Decimal" or Library of Congress, or perhaps ISBN?
• Is the "URL" somewhere the user can go to a) buy a copy of this work, or b) see an electronic version?
• Is "Year Range" intended to denote a range of time covered by the work? A range of publication dates? or perhaps the DOB-DOD range of the author?

It would be possible to craft a bibliographic citation automatically from the entries in these input boxes, but you'd have to match up the input boxes with the information actually needed for the citation. That is, for a book source there'd be a need for

• Author
• Date
• Title
• Place of Publication
• Publisher
• Pagination

and there would need to be a different suite of data for other types of sources (e.g., family bibles, manuscripts, journal articles, etc.

Bill 09:46, 9 November 2007 (EST)

Hi Bill,

I get confused by the label, type. Are you referring to type of record, i.e. deed, probate, vitals, census or the format of the record, i. e., book, online database, digital image, letter, email, message board? --Beth 09:34, 27 January 2008 (EST)

I don't know. Those are the titles of the input boxes that come up, but I'm not sure exactly what the programmer meant them to hold. Bill 21:26, 28 January 2008 (EST)

Proposal and Questions [13 March 2008]

First of all thank you to everyone for the great discussion on this topic. It's so nice to get the benefit of everyone's experience. With the launch of the digital library in the next couple of months I realized that it would be very helpful to nail down the new source format before launch, so I've spent a lot of the past week reviewing this document, Elizabeth Shown Mills' book, and the Dublin Core (a set of fields developed by the library community for describing items in a digital library). Based upon all of this, I'd like to ask for feedback on the following proposal. I imagine I don't have as much experience citing sources as most of you, so I really do appreciate the feedback. I don't expect to get to cleaning up the existing sources until after match+merge, but we can at least start gathering the correct fields for new sources and repositories now.

First, I think it's helpful to decide what we want the ultimate goals of the source wiki to be. Here are the goals I've come up with:

1. Discovery: helping people find sources that they can look up for their research
2. Encourage good citation by making sources easy for others to re-use: one reason most people don't source their research is because creating the source records is daunting and time-consuming. A ready-made set of existing source records would help more people add sources.
3. Provide a place to add usage notes and repository information - anything that would help someone else find and understand how to use the source.
4. Someday, automatically-generate source citations in printed family histories based upon the information in the source citation on the person/family pages and in the source pages. After spending a fair amount of time looking at ESM's book, I don't think we'll be able to duplicate her exact format in all cases because it varies too much, but I think we can get close.
5. Provide a place for source citation information for items uploaded into the digital library.

Proposal

We classify sources into 7 types. When adding a new source, you

1. first enter is the type of source.
2. Based upon the type, you're asked to enter a few additional fields so that the system can generate the title of the wiki page for the source (see the table below).
3. Next, the system displays a list of existing source pages with similar titles (if any) so that you can choose one of them if they're what you're looking for instead of creating a new source page.
4. When editing the source page, the fields you're asked to enter differ depending upon the type of source (see the second table below).

Fields to gather on Source pages [20 March 2008]

In addition, the following fields would be gathered for all types:

• Repository (can list multiple)

• Name - title of a Repository page
• Location (call number, film number, or URL)

• Surname(s) covered -- major surnames covered by the source
• Place(s) covered -- places covered by the source
• Year range covered -- years covered by the source
• Ethnicity/Culture covered -- if source pertains to a specific ethnicity/culture (we'll have a drop-down list of 50-100)
• Religion covered -- if source pertains to a specific religion (we'll have a drop-down list of 50-100)
• Occupation covered -- if source pertains to a particular occupation (we'll have a drop-down list of 100-150)
• Source category -- e.g., Vital records, Census records, Biography, Finding aid, Map (we'll have a drop-down list of about 30)

NOTES

• The decision of whether or not to create a new source page for sources based upon each other is whether or not the source has a different title and author. Books that have been printed several times could have multiple publication dates.
• The reason for listing authors as "surname, given names" is using this form we can generate the names in "given names surname" format if we need to for citations. It's more difficult to go the other way from "give names surname" to "surname, given names" format when the surname contains a space (e.g., Linus Van Pelt).
• Note that the following fields currently on source pages will not be gathered on the new source pages:

• Call number -- If this field is present, I'll add a repository with a blank name to the source and move this field into the repository location
• URL -- similar to call number
• Repository Address -- should be tracked on the Repository pages
• Publication info -- tracked as three separate fields now: publisher, place, and date. Existing "publication info" field values will be moved into the publisher field for lack of a better place to put them. If they're of the form "X: Y, year" then I'll move Y into the publisher field, X into the publication place, and year into the publication date.
• Type -- according to the GEDCOM spec, Type is meant to contain what I would call "format" -- whether the source is audio, book, card, electronic, fiche, film, magazine, manuscript, map, newspaper, photo, tombstone, or video. The problem with recording type in this sense is that in our definition of source, a single source might be several types (e.g., a book that has been microfilmed). A different type could be associated with each repository, like the Repository Location field. So I could add "Repository Format" as a field alongside Repository Location, but how helpful is it to track type?

Fields to gather on Repository pages

I'll create a new Repository namespace with the following fields:

• Place - city or town -- links to a place in the place wiki
• Complete postal address
• Phone number
• URL

Fields to gather in Source citations on Person/Family pages [26 March 2008]

We're already have fields for the following:

• volume/pages/dwelling -- volume number of book, page number or range, and/or census dwelling and family information
• date issued/accessed -- date specific printing of the book you used was published, date of record, or date on-line information was accessed
• quality -- unreliable, questionable, secondary, or primary -- from GEDCOM specification
• description

We'd need to add the following fields when the citation is to a Source page:

• Record name -- name of the specific record you looked at (e.g., John Smith birth certificate #10256)
• Repository and location/URL -- name of the specific repository that you found the record in, when the record is available in multiple repositories. Also includes the location of the record in the repository (e.g., roll# or URL).

I suppose I understand why the repository field is listed, since technically a proper citation includes exactly where a person found the information. But I think this will quickly become a mess in a wiki environment because the point is no longer where a single person found the information, but rather where the information can be found. At that point, such information will just confirm the obvious -- that a given page is actually in the book at X library. Of course it is (except in really bizarre circumstances that can be noted on the source page). And roll numbers or URLs belonging to a repository belong on the source page, it seems to me, since they'll be consistent for every person that uses the source at that repository. So we're getting exactly no useful information in return for a big messy citation record where more than a couple people have chimed in.--Amelia 23:09, 13 March 2008 (EDT)

Yes, that's the big question: should we store "Repository+location" in the source citation. If our Source's are found in multiple repositories, and if we want to someday generate "proper" citations from wiki pages, then we have to store the particular repository someone found the source in along with the exact roll# in the source citation. I'm ok with leaving it out (although some people have asked for URL and we might want to leave that), but I'd like others' opinions.--Dallan 16:26, 14 March 2008 (EDT)

Don't forget to plan for success, Dallan :-) The only way you get "proper" citations from this design is if only one person enters where they found the information -- and then the citation is only "proper" for that person who used that repository. For properly wiki-fied entries where a dozen people collaborate, you'd end up with an unwieldy list of places the info has been found, and it's not technically "proper" for anyone who didn't look at every place on the list. To put it another way (since I know the conception I'm arguing for contradicts what a lot of people think is proper), such specific detail seems to have two purposes: documentation for the person who did it, and direction for someone looking for the same information. The documentation part doesn't belong on a wiki entry because it doesn't "belong" to any one person saying "this is what I did". The direction part can be accomplished by the (hopefully) more complete repository information on the source page. I will say that in situations where the source is not especially specific (like if we do "1880 census"), then roll # or URL should be on the person page cite, but that's not the same as repository.--Amelia 22:21, 14 March 2008 (EDT)

Any other thoughts? If we're just going to track Roll# / URL, we could probably store that in the "volume/pages/dwelling" field (and maybe generalize the name of that field to "Location within source"), and omit the "Repository+location" field.--Dallan 15:47, 17 March 2008 (EDT)

I think Amelia is right -- having the repository for something like a book could be problematic. However, some books are rare and having the repository info would be helpful. Thinking down the road, would it be possible to auto-generate a link to a book's entry in WorldCat.org? This would help people find those books, such as some family histories, where there might be only a dozen or so in existence. Lots of libraries have History of Fairfield and Perry Counties, Ohio (so listing the repository would be messy and probably pointless), but very few have The Dildine Family in America. The WorldCat.org link would help people find it. --Ajcrow 10:49, 23 March 2008 (EDT)

We ought to be able to do that. In fact, we could automatically add a "Search on WorldCat" link to every Book source page. I'll keep that in mind.--Dallan 11:22, 25 March 2008 (EDT)

I use the WorldCat quite a bit myself. Its a great tool. It does have a failing though, in that it does not always get the citation right. I don't know if the citation creation is something being driven by a program of some sort, or if this is manual human entry. In either case though, I've notice that a small number of citations contained on WorldCat, are not at all the way I'd prepare the Citation. For example, if its a Journal article that they are citing, there's a tendency to use either the organization publishing the Journal as the author, or to cite an editor as the author. An article written, for example, by John Smith and published in the New England Journal of Genealogy might get a citation listing "The New England Genealogical Association" as the author. (Just made that one up, but its similar to instances I've seen. Using World Cat as the standard would be good, I think, since these errors, while not exactly rare, are not common. There is a need, I think, to be able to over-ride the WorldCat entry if that makes sense. Q 12:14, 25 March 2008 (EDT)

I'm thinking that we'd have a "Search WorldCat" link displayed on Book (and maybe Article?) Source pages. We wouldn't automatically copy any citation information from them.--Dallan 11:55, 26 March 2008 (EDT)

MySource pages [22 March 2008]

I'm thinking that in the future, MySource pages will generally be created only during GEDCOM uploads. We would discourage their creation on-line. All sources created on-line should use the Source namespace. Eventually I'll try to match sources in GEDCOM files to existing Source pages, but when a good match can't be found, we'll still need to create MySource pages for them during GEDCOM uploads.

NOTES

• Should we rename MySource to GedSource? This would require some work, but if we decide that MySource's should only really be used by GEDCOM uploads, and that all sources created on-line should be Source pages, do we want to rename the MySource namespace to GedSource to reflect that fact? Or would GedSource connote "this source refers to a GEDCOM file" (which is incorrect) more than "this is a source that was generated during a GEDCOM upload" (which I think is what we want)?

Question Should MySource pages be used primarily just for sources created during GEDCOM uploads? What about not-very-reusable sources that people want to cite, such as:

• letters, journals - especially when the user has not uploaded scans of the pages as Image pages or into the digital library
• family trees - especially when the user has not uploaded a GEDCOM of the tree into the digital library
• engravings on rings, bracelets, etc.

Should we encourage people to create MySource pages for these things, in which case the definition of MySource becomes sources that are not re-usable -- that pertain just to your individual genealogy, or should we encourage people to use "title only" source citations for these, in which case they do not appear on a source page at all - just in the source citation on the individual person or family page. In the latter case we could say that MySource's are really only for GEDCOM uploads, and that non-re-usable sources should be title-only sources. Thoughts? Is "MySource = source that is not re-usable" a simple enough definition to distinguish MySource's from Source's? There's been a lot of confusion in this area, and I'd like to come up with a definition of MySource that puts the confusion to rest.--Dallan 16:26, 14 March 2008 (EDT)

I vote for 1) not renaming - I think it connotes a gedcom source, which is often not the case; and 2) continuing to permit MySource to be used for limited-use sources. For example, I have a family bible entry and a pension file family group chart as MySources. The editable nature of the page lets me put all the detail in and cite it a half dozen times, and provides that handy "what links here" feature. While those could be Source pages, they're not really of the same general interest and segregating them makes some sense, but having the information all on one page saves a decent amount of work. Since that savings comes because they are actually reusable in a limited sense, perhaps a definition like "MySource = Source that other people are unlikely to want to use"--Amelia 22:28, 14 March 2008 (EDT)

I continue to like the concept of "MySource" being restricted to GedCom downloads. I've used it for my own purposes simply because there wasn't a category where I could comfortably place "snippets". The digital library will probably serve that purpose in the future. The real advantage of "MySource" is that it creates a place for the ephemeral sources used by the vast majority of web genealogists. Thus it meets a real need, but at the same time provides a quiet nudge that other sources might serve the genealogical purpose better. Bill 08:50, 15 March 2008 (EDT)

I think that it would be beneficial to segregate the sources from GEDCOMs, but I also think that we should separate out sources of limited use to the general public. Why can't we do both. Keep MySource's and rename GEDCOM upload sources as something else. This would make it easier to discern what is a source, what is something that likely someone hasn't looked at since uploading their GEDCOM, and what is something that is really only important to a few people.--Jlchaff 23:35, 15 March 2008 (EDT)

I don't want to create a third namespace because we already get usability complaints that we have too many different types of pages, and here we're going to be adding two more: Repository and Digital library item. Based on the discussion, I'm leaning toward not renaming them and encouraging people to use them for limited-use sources (as Amelia suggests) and ephemeral sources (as Bill suggests) as well as sources from uploaded GEDCOMs. In order to be consistent with User pages, I'm thinking that from now on the title of any MySource page you create automatically starts with your user name, and you're the only person who can edit it , although I could very easily be convinced that we should allow others to edit your MySource pages as well. What do you think?--Dallan 15:44, 17 March 2008 (EDT)

Does the MySource page somehow take up less space than a Source page on the wiki; is this why you recommend My Source for limited-use sources? --Beth 09:31, 22 March 2008 (EDT)

No, the reason is to help with the goal of resource discovery -- in the future we'll allow people to search Source pages separately from MySource pages, and when people search or browse Source pages, it would be nice to show primarily sources that were broadly or semi-broadly applicable, without the result list being cluttered too much with poorly-specified sources from GEDCOM files or limited-use sources that others wouldn't care about or have access to.--Dallan 16:43, 22 March 2008 (EDT)

The only reason I've edited someone else's MySource is to redirect it to the actual Source page (for something out of a gedcom). The same thing could be accomplished by telling the user to do it himself, but as a practical matter, that's often unlikely to happen, so it's nice usability-wise to just be able to fix it.--Amelia 16:46, 17 March 2008 (EDT)

I believe that others should be allowed to edit a MySource page. Some users may upload data here and not edit pages ever. There is also always the possiblity of the death of the user; for these reasons I think most pages should be editable by any user.--Beth 09:31, 22 March 2008 (EDT)

Ok, let's allow others to edit. Especially if we say that limited-use sources should generally be MySource's, I can see that it would be possible for several family members to want to all refer to, and possibly edit, the same letter/journal/ring engraving MySource.--Dallan 16:43, 22 March 2008 (EDT)

If the purpose that is truely intended is to capture those very personal sources that so many like to use (GEDCOM, FamilyTree, email), then I wouldn't think anyone but the user would care what the thing said---its solely for their convenience. From that perspective, no one else would ever have a need to modify the entry. Indeed, that might run counter to the interests of the user. Now, in my case, since I'm not using the name space in quite the way it was intended, I'll probably have to make some changes---probably sending things to the digital library, being the obvious answer. Hopefully someone will wave a magic electronic wand, and that will occur automatically, but perhaps not. In which case I'll be toting and lifting stuff for a bit. But that's OK. I'll pay for my own idiosyncracy.

And on a related point---making the system EASIER to use, is probably very important. Every system has its owns quirks, and fancies, and sometimes those systems get overloaded with very clever bells and whistles designed to meet some very specific needs. Then you end up with something like Microsoft Word, where the bells and whistles become some abundant that the usability gets degraded. I'm fairly comfortable with Word, but dang! Sometimes I wish it would just let me type, instead of trying to figure out what I'm doing, and doing something I don't want. And sometimes its a pain to wade through the bells and whistles, just to figure out how to make it STOP doing something. Keeping it simple is a good idea. Q 18:09, 17 March 2008 (EDT)

Further Discussion [25 March 2008]

• Censuses: A lot of people cite censuses by county, and I can see there's some benefit to that. But realistically, it's a lot of work to create separate source pages for each county for each census. I don't know if I would prohibit someone from creating a county-level source page, but I wouldn't want to make the effort to create them myself.

There is a lot of information in this document, and I'm sure I've missed something. Feel free to bring up additional points or questions.--Dallan 16:04, 13 March 2008 (EDT)

I'm no fan of ancestry.com, but the one thing that I think they get right is a very nice census reference. It contains the actual government retrieval information (film reels and such) as well as information in two associated notes that lets you actually jump to their copy of the image (if you've got a subscription). See the sources for the 1910, 1920, and 1930 US census on Person:James Mason (13) for examples. I don't know how many of those references I've got - it looks like about 250 for the 1930 census alone. All I would really want to do is turn those per-year sources from "mysource" to a well-documented regular "source".--Jrm03063 19:15, 13 March 2008 (EDT)

Ancestry's citation for the 1930 census is

Online publication - Provo, UT, USA: The Generations Network, Inc., 2002.Original data - United States of America, Bureau of the Census. Fifteenth Census of the United States, 1930. Washington, D.C.: National Archives and Records Administration, 1930. T626

If we had a single source for the 1930 census, we could someday automatically generate the middle part of this citation:

United States of America, Bureau of the Census. Fifteenth Census of the United States, 1930.

This source would show that the census is available in multiple repositories: Ancestry, NARA, etc. If we store "Repository+location" as part of the citation (see a discussion about this above), then you could enter the following in the "Repository+location" field

Online publication - Provo, UT, USA: The Generations Network, Inc., 2002, citing Washington, D.C.: National Archives and Records Administration, 1930. T626

and someday we could automatically generate the citation as:

United States of America, Bureau of the Census. Fifteenth Census of the United States, 1930. Online publication - Provo, UT, USA: The Generations Network, Inc., 2002, citing Washington, D.C.: National Archives and Records Administration, 1930. T626

But I'm not sure how to automatically generate a citation using the exact form Ancestry's using without either (a) creating a separate source for the Ancestry version of the census, or (b) storing a separate "Citation" field, where you could enter the citation exactly how you want.--Dallan 16:26, 14 March 2008 (EDT)

Hi Dallan; I understand why you do not wish to create the census source at the county level; but perhaps you could allow one to create a subsource; that would give users a choice and the source would be available at county levels if a user chooses to create one. --Beth 20:21, 13 March 2008 (EDT)

That's a possibility, although regular source pages that link to (cite) the main source page might be simpler than subpages. I'm not against county-level census sources, I just don't want to create them myself :-). I could be convinced either way on this.--Dallan 16:26, 14 March 2008 (EDT)

I've heard noises about a public domain source for census record images in the future. Until that's available, what you need in a census record citation is the information on the census page, enumeration district, film roll, etc. Essentially, all the stuff that you would need if you went into the national archives to look at film, or went to a library where they have it via ancestry. When there's a public domain source for the images, we'll probably want to cite the image, but that's not here yet.--Jrm03063 21:12, 13 March 2008 (EDT)

I'm thinking that the "Volume/pages/dwelling" field in the source citation would be where you would enter the enumeration district and dwelling number, and that the "Repository+location" field in the source citation would be where you would enter the film roll. Does this make sense? Alternatively, we could ask people to put the film roll also in the "Volume/pages/dwelling" field, although doing so would make it more difficult to generate a "proper" citation someday (see discussion about Repository+location above).

I'm confused on this point. Where will the state and county go in the source? "United States of America, Bureau of the Census. Fifteenth Census of the United States, 1930." doesn't tell me enough for the source to be of use. Even with the addition of the roll number, the state and county aren't immediately evident. --Ajcrow 10:56, 23 March 2008 (EDT)

The user would still have to type the county/city into the citation detail on the family/person page (under Record Name, I guess); the source page title isn't necessarily meant to be a full citation, it's just a link to further information on the source. (Which does raise the question -- if there were only source pages for each census year, I don't think I'd even bother linking to them; the usability gain from the link would be offset by its length.) --Amelia 11:10, 23 March 2008 (EDT)

And we'll allow people to create county-level source pages if they want to.--Dallan 11:22, 25 March 2008 (EDT)

A question, and an Observation [23 March 2008]

Dallan, after looking over your proposal, I see many things I like. I also see some things that won't work particular well for me, but I suppose I will find workarounds. However, I do have a question or two that I'd like to see your response to so that I'll understand the shape of things a bit better:

First, What is it that you consider to be a source? Is it the actual information, or is it a pointer to the item in question? For example, Source:Scudder, 1899 contains considerable information.

• It could be extracted and placed on an article page (or many pages) and referred to as a "Source".
• On the other hand, it could be simply a bibliographic citation (e.g, as Source:Scudder, 1899.

The term "source" is often used in both of those senses, but it would be helpful to understand how you are using it. While you can use the word in either sense, you can't do so simultaneously without creating some confusion.

I'm thinking that we should say that Source pages contain "meta-data" about the source, and not the source itself. So they would contain citation information, coverage information, usage tips, but not actual data - your second definition. If someone wants to include an actual source, they'll upload a bunch of scanned images and/or a transcription into the digital library and the digital library item will link to the source page (and the source page will automatically link to the digital library items that reference it). Or they can create an article with the information and link from the Source page to the article. I agree this has been a point of confusion; I'm hoping that this definition will make things less confusing in the future.--Dallan 16:26, 14 March 2008 (EDT)

Thank you. Clarification much appreciated.

Second, (the observation) I notice you include "Web pages", "message boards" and "family trees" in your list of sources. I know lots of people use these as their "sources". I myself will occassionally point to one of them in a temporary form, simply because I'm using them as I gather information and may want to get back to them to check something. But long term, I would not use them as a source---not because the quality of information is poor--that's sometimes true, but often these sources have excellent information. (That's why I use them when I'm in information gathering mode.) The reason I wouldn't use them as sources is that for most of them (especially personal web sites) there is no assurance that the information you found there today, will be there tomorrow. The reason we cite sources is because those who read our work need to be able to go back and see that our interpretation of a source is correct. To do that, they have to know what we used as a source. Therefore we have to tell them where to go to find the information we are using. If we don't do that they can't verify the information, and we can't show that we have correctly interpreted family relationships.

All web sites (including Ancestry, message boards, and for that matter "weRelate", are ephemeral. Not only is there no guarantee that the site will still be operational at an indefinite time in the future, there's no guarantee that the information will be the same if the site is still around--say 20 years from now. While I'm sure everyone understands how personal websites have a very finite existence, I'm not sure many people realize how ephemeral message boards and the like are. I can name several instances where significant amounts of work have been removed from specific message boards and email archives (quite literally hundreds and thousands of messages). How much use is it to cite a Family Tree posted on Ancestry, when the content of that tree changes on a daily basis?

So, if the objective is to get people to cite things well, is it a good idea to suggest that they cite ephemeral sources? Bill 20:51, 13 March 2008 (EDT)

At the end of the day, we have to rely on the community to develop a passion for quality research. There are good sites of lasting value - wikipedia for one. In the future, there will probably be others. We have to hope the community will develop a passion for quality research. When I see a weak citation that rubs me the wrong way, I don't junk it, I try to dig up a better one and put it in along side.--Jrm03063 21:44, 13 March 2008 (EDT)

A good example of "nothing is written in stone"

a letter from Tim Sullivan CEO The Generations Network, Inc.:

we have decided to "transplant" RootsWeb onto the Ancestry.com domain beginning next week. This move will not change the RootsWeb experience or alter the ease of navigation to or within RootsWeb. RootsWeb will remain a free online experience. What will be different is that the Web address for all RootsWeb pages will change from www.rootsweb.com to www.rootsweb.ancestry.com. Again, the RootsWeb experience is not changing

I believe that somethings are in fact written in stone, but of course, even stone weathers away---visit any graveyard to see the proof of that. Still, you can expect to be able to revisit some sources today, tomorrow, or a hundred years from now. Libraries are very good at preserving perishable materials, but of course there's always some lost with time. The trick is to find sources that have a reasonable shot of preservation. Personal websites, for example, are the most ephemeral of sources; people often point to them as their source, but of course in fairly short order, when someone else seeks the same information, the site is no longer on line, or has otherwise changed so that the information can not be recovered again. Sites like Ancestry are somewhat better than this, as you can have some reasonable expectation of revisiting that particular site in a few years, and finding those now ancient messages. Ten years from now I'm not so sure about---and 100 years boggles the mind. But I bet if I cite some published work that I've found in a library (or even on Google Books), someone else will be able to go back to that same work 100 years from now, and find the same thing again---if I tell them where I found the information. The technological world does evolve, and perhaps in ten years, capacity will be so immense that virtually nothing is ever lost and can always be recovered. We aren't there right now, and I don't know that we ever will be. Which is what makes citing ephemeral sources a not very good idea.

And while I'm on that subject, I'm fairly confident that this site will evolve into something quite different over the next ten years---assuming it doesn't go the way of stone. I don't know, of course, what it will turn into---but I'm sure with technological innovations being such as they are, it will evolve, and transform itself. I'm hoping the site itself will stick around long enough for me to complete my principle objectives. But right now, who knows in what direction it will take? Bill 13:04, 14 March 2008 (EDT)

I'm ok with deleting the sources we have for individual websites unless someone is already linking to them. I don't think they add much to the source index. I could also be talked into deleting the message board sources, since we have so many (250,000) of them and I bet most of them will never be cited. So then it's a question of whether we allow (which might be viewed as encourage) people to create new Source pages for message boards or websites, or tell them to create MySource or Title-only citations for them. I think I'd rather help people document even poor practices by allowing them to create Source pages for these things than not. What do others think?--Dallan 16:26, 14 March 2008 (EDT)

ultimately, I think the criteria for inclusion or exclusion, is probably whether something can be revisited long term by others..."to see the same as I", as it were. That's something of a value judgement, or prediction of survivability. My personal preference would be not to include things like message boards and mailing list archives, simply because they are, at their core, ephemeral. Bill 16:43, 14 March 2008 (EDT)

I think I should clarify something. I'm not in favor of telling people they can't cite whatever it is they want to cite. Good source, bad source, permanent source, ephemeral source---whatever it is they are comfortable with citing, is what they should cite. The reason for that its better for people to cite an ephemeral source, than it is to cite no source. What I'm in favor of is encouraging people to make use of non-ephemeral sources. I like the MySource concept as laid out by Dallan, precisely because it allows people to cite whatever it is that they are comfortable with. What I was suggesting was that "Sources" not include options for message boards and the like. There is a need, I think, for "Sources" to be restricted to works for which there is some potential for them to be revisited today, tormorrow and for the indefinite future. Q 08:21, 21 March 2008 (EDT)

If we encourage people to use MySources for "ephemeral" sources, we could do away with the "Message board / Website" source type. Most of these would generally be MySource pages, and people could use the "Miscellaneous" source type to create a Source for a website if it was really appropriate.

I'd like to keep "Family Tree" as a source type, but tell people that it should be used just for trees that are being uploaded into the digital library so the tree will be accessible to others in the future. MySource pages should be created for other trees.

What do others think about this?--Dallan 15:44, 17 March 2008 (EDT)

I'm the first one to rail about poorly sourced trees where people just cite other people's web trees, but I'm also a bit concerned about tarnishing all websites with the same "ephemeral=bad" brush. There are actually well-researched, useful, unique sites out there that deserve to be treated with the same respect as other sources. The weight of a source depends on what goes into it, not the medium. Of course someone could pull the plug, but the library holding the only copy of that manuscript could burn down too -- that's not a reason not to trust or cite the source.

On the trees, I'm happy with whatever method of discouraging citation to them you want to come up with.--Amelia 16:42, 17 March 2008 (EDT)

That's a good point. I think we can accommodate websites under the "miscellaneous" type. If the user chooses "miscellaneous", we'll have to show a message outlining the types of websites that it is appropriate to create Source pages for.--Dallan 01:38, 20 March 2008 (EDT)

I believe Amelia misconstrued the point. There was no equating "personal web site=ephemeral site=bad site". Many personal web sites are are both information rich, and the conclusions well researched. Some are not, but the quality of personal web sites is not the point. The point is whether its useful to cite a source that has a relatively short half life. Yes, libraries can burn, but not all libraries will burn. When you publish a work, consign it to paper, multiple copies are created. They'll have some limited distribution. The work Source:Scudder, 1899 was originally published in exactly 500 copies. Doubt that it was ever republished. Yet 109 years later its still available, and quite useful for work on the families of Huntington, Long Island. I doubt that all 500 copies still exist, but multiple copies still exist because you can find a list of libraries that house this work. More critically, a copy made its way to the scanner of Google Books, and is now electronically preserved. Perhaps not forever, but on a formal site like Google there's some assurance that the work can be consulted for an indefinite time in the future.

In contrast, when you "publish" something on your personal web page, you'll probably get much broader distribution immediately. After all, virtually anyone from the banker in New York, to the farmer in Angola, can potentially get to your work. But what's there is only going to be there as long as the author chooses to keep paying their service provider. More over, there's no guarantee that what's on that page today, will be there tomorrow. May be very similar, but the author might also make changes in that page. Today they might say that John Smith's DOB was 1 October 1878. Tomorrow, they might get new information, and decide his DOB was really 1 October 1877. Possibly they will document that, possibly not. But in anycase, once the page has changed, it can't be revisited (at least not easily). And if it can't be revisited, you can't verify or validate what you once found there. Hence, personal web sites do not meet the criteria of a verifiable source. And sources that can't be verified are sources that I believe should be avoided.

One might also add that WeRelate itself, by its very nature, varies on a daily basis. The advantage here (and on other wiki's) is that the history of the changes to any given page are available for viewing. So (if you wanted to cite a specific WeRelate page for 21 March 2008 as a source, you could in theory do so with the expectation that others come back to exactly that same page and see what you saw. You could do that today, tomorrow, or a year from now---as long as you specified the exact date of the page that you were citing. Of course, the value of citing a WeRelate page depends not just on the quality of the page, and its internal documentation, but also on the long term persistance of the WeRelate site. I trust and hope WeRelate will be around for a long time, but things happen. The history of Rootsweb is a case in point. Whatever you might think of the liklihood of Rootsweb persisting long term, its quite clear from current and past traffic on certain email lists that more than a few people have a concern over their ability to reach material stored on Rootsweb sites in the long term.

Q 08:10, 21 March 2008 (EDT)

From the discussions I don't think we're going to be able to develop hard-and-fast rules for when a Source or MySource page is more appropriate. I think we can display something like the following as guidelines.

Source pages should be created for sources that:

• are publicly available for others to look up and verify,
• contain citation information so others can locate them,
• are expected to be accessible years from now, and
• apply to multiple individuals.

Consider creating MySource pages for sources that do not generally meet these guidelines.

Some people will interpret these guidelines more strictly than others, and I think we need to be liberal in allowing a range of interpretations especially at the beginning.

There's another reason though to remove "Website / message board" from the list of types, which is that it also removes some confusion for the user who wants to create a source for a record set they found on a website; e.g., a census they found on Ancestry. If we have "website / message board" in the list, some people are likely to create a source for the website (e.g., Ancestry.com) instead of creating a source for the record set (e.g., the census). That's what finally tipped the scales for me -- to remove this possible point of confusion. We can add verbiage to the "Miscellaneous" type explaining that it is appropriate to use this type for websites / message boards that meet the above guidelines. We'll also ask people to be careful not to create Source pages for websites that are being used as repositories of records found in offline sources.

How do these guidelines sound; any suggestions for modifying or re-wording?--Dallan 16:43, 22 March 2008 (EDT)

I think you've hit upon a good solution. While some websites, message boards, etc. are not the best source for information (then again, some books are pretty lousy, too), they are still sources. If that's where I got my information from, I should have the ability to say that's where it came from. People looking at my research can see the citation (even if they can no longer visit the actual source) and decide for themselves if they think further research for that point is warranted. --Ajcrow 11:03, 23 March 2008 (EDT)

(I know this is a bit off-topic, but for preservation I'm a believer in the concept of Lots of Copies Keeps Stuff Safe. To use this approach, it's a huge help to have an open-content license that allows others to freely download your material. They become your company-external backup. If the Wikimedia Foundation were to shut down completely tomorrow, we'd be able to find a copy of Wikipedia that was just a few days old from one of the many people that download and use the Wikipedia database. In addition to multiple internal copies, LOCKSS is a direction I want to pursue for both the wiki and digital library.)

Sources [22 March 2008]

I'm breaking this up into smaller chunks as the discussion is getting fairly complex.

The Problem [23 March 2008]

Discussion moved from WeRelate talk:Watercooler--Dallan 09:43, 21 March 2008 (EDT)

I'd like to get some input on sources.

Once upon a time I created an article at

MySource:Will of Thomas Scudder (1657/58)

This item was obtained in response to a GenForum message asking for the will of Thomas Scudder who died sometime between Sept 1657, and June 1658 (will and probate dates), in Salem Mass.

I placed it in MySource namespace because I consider Message Boards to be an ephemeral source---ones that I can't guarantee someone else can ever get back to. And if no one can get back to it at some indefinite time in the future, its not useful to cite it. In truth, this applies more to personal web pages than to a long term site like GenForum, but citing GenForum as a formal source is probably not really a good idea on general principles. At best its a tertiary source.

Fortunately, the message provided the original source in Essex County records, which (if I had seen that) would be what I could cite as the source. But I haven't seen those original records, and am not likely to, so this is not the right answer as to what to cite.

I understand that this will has been published in Essex Antiquarian , a genealogical journal for the area. I've got a complete citation for that, (see the MySoruce article for the cite), but again, this is not something I've actually seen. (But I could realistically get this, or a photocopy of the article, and let it go at that.) But that will take time.

In the meantime, I'd like to be able to place a reasonable source citation on this record. (This is related to work on the Digital Library now in alpha testing). As an aside, to get this will into the digital library I need an article in Source namespace (not MySource), to pin it to. So I created a Source article at

Source:Will of Thomas Scudder, 1657/58

And that brought up a whole lot of other questions. Things I've noticed before, but ignored because "Source" concept was in development. Since I'm working on the digital library, I need a better handle on these fields, so I'd like to see how others view them.

The data entry page asks for various pieces of information when creating a source. Here's my list of questions.

Surnames covered(one per line):

-- This is fairly obvious. "Scudder", or if I get ambitious, the surnames of everyone listed in the will.

Places covered (one per line):

---also obvious, Salem, Essex County Mass.

Year range:

- Not quite so obvious.  The will was written in 1657, but entered in the court records in 1658.  Perhaps the "range" is 1657-1658, but I'm not sure that's what is intended.

I recommend using the date of the will because you are citing the will, not the filing date or recorded date. --Beth 10:18, 21 March 2008 (EDT)

I agree; the possibility of a year range is for collections that cover a range of years.--Dallan 16:43, 22 March 2008 (EDT)

URL:

---is this supposed to be a link to an electronic source for this item? Working on the digital library makes me suspect that its really the address of this article, but I'm not sure.

I question citing the URL because URLs change through the years. Who has the time to fix all of the eventual broken links? --Beth 10:18, 21 March 2008 (EDT)

I agree in general. If the link is to the source, and the source is ephemeral its going to break very quickly (in a meaningful genealogical timeframe). However, if the source is something in say "Google Books" that we can expect to be around more or less indefinitely, than this would not be a problem. Linking to a personal website would be a problem. A GenForum message probably falls somewhere intermediate.

However, the real issue is what was INTENDED to go into this field? A link to a WeRelate article, or a link to the source where the item was obtained, or a link to a source where someone else might reasonably expect to find a copy at some indefinite time in the future. Q 09:34, 22 March 2008 (EDT)

What I had intended to go into this field is the URL of an online source. In the new source format, we're going to move URL down to the repository section, since this is really the location of the source within an online repository. (Sources will be able to link to multiple repositories, some offline, some online in the new format.)--Dallan 16:43, 22 March 2008 (EDT)

Intuitive labels are very good things. The more intuitive the better, and the best label is one where no one has to think at all about what it means. "URL" could mean a lot of things, so perhaps what's needed is a label that is more intuitively obvious. Something like "Electronic Source" could be used in place of "URL" when this is the sense of what the field is supposed to contain. Q 11:29, 23 March 2008 (EDT)

Author: [7 June 2008]

Who would that be?  The person making the will?  The witnesses?  The Court clerk who recorded the will?

The author is the person who wrote the message on Genforum. If you were citing the record itself, the author is generally the government entity that created the record. --Beth 10:18, 21 March 2008 (EDT)

Possibly, that's one of the issues here. This hinges on what is perceived as the source of the information: The original will, the microfilm record that someone found the will on, the Essex Antiquarian, or the message board answer, or the person who wrote the message board answer. Q 09:34, 22 March 2008 (EDT)

This is an area where genealogists get hung up. A professional genealogist will always tell you to cite exactly where you got your information---even if its an ephemeral source. That might be a good answer (professional genealogists would probably say its the only acceptable answer), if your purpose is simply to be able to retrace your own footsteps exactly, and in the case of ephemeral sources, and your problem is only a short term one ((e.g., you expect to get a better (non-ephemeral) source soon)). In that case you can probably live with the potential that your short term ephemeral source will evaporate. But here a short term goal is probably not acceptable. There's a need to point to a source that is verifiable in the long term. Possibly the Message Board source would do, but Essex Antiquarian or the Court records, serve that purpose better. That fact that I haven't seen those sources is a downside, but since this IS a social networking site, perhaps someone else will note this, have immediate access to one or the, and confirm that the will as written is actually in one or the other of those sources. Not what the professional genealogist would tell you to do, but maybe this is a problem professional genealogists haven't really had to deal with in the past. New technology, new problems, new solutions. Q 11:29, 23 March 2008 (EDT)

Societies as Authors [10 June 2008]

How are we handling publications put out by Societies, such as the DAR for example? Do we put the society name in the author field such as, Daughters of the American Revolution (Indiana). A Roster of Revolutionary Ancestors ? --Ronni 23:43, 7 June 2008 (EDT)

I've noticed that some cataloguing sources (ie, WorldCat) will cite the society, if there name appears on the Title page as having sponsored the effort, even when there's a legitimate author. The fact that DAR might have sponsored something doesn't seem to me to have a great deal of relevance. I'd rather know who actually did the work (the person), rather than the sponsoring organization. The exception to that is when the work is a cooperative effort, in which many people participated, and many people deserve credit. I can see that happening in a DAR sponsored project, where lots of folks contributed individual paragraphs, and no one person should receive the credit owing to the many. So the solution is simply to say its a DAR publication, and get on with it. But often, the exact role of the DAR or other organization is unclear. They appear on the title page, but its not obvious what contribution, if any, they actually made. In those cases, I usually ignore the organization,or place it toward the end of the citation, but not as the author.

This is somewhat analogous to what happens with some governmental publications. A work is prepared by staff of some agency, and the agency head gets his name prominently displayed on the title page. Those who actually wrote the thing appear in the "List of preparer's" in an appendix in the back. In those cases, the citation usually reads somethng like "US Department of Energy, 2006", but sometimes, especially in the popular press if it gets picked up in a newstory, it will be cited as "Smith, 2006", where "Smith" is the agency head. He probably never saw the work, or if he did, his contact with it was to read the abstract prepared by "the list of preparers".

But in non-answer to your question, what do you think "the rule" should be? And should there be a rule for something like this? Q 07:56, 8 June 2008 (EDT)

As to a rule for this, because I lean more towards the aesthetic - it has to look good first, then I worry about its functionality and/or practicality - I would leave off the society name. I think society names would get unruly. Secondly, I think the main rule of thumb for me personally on this would be: if I don't have a SURNAME, GIVEN NAME as seen on the cover or title page of the source to enter as the WeRelate fields indicate, then I'm not going to worry about it as far as titling goes. All of that extra info can be added to the page contents. The title should be clean and simple, but yet informative as to what the source is, in my opinion. --Ronni 12:09, 8 June 2008 (EDT)

Usually what happens is that the authors name IS on the title page, its just that the Organization likes to slap their name on that page as well, and sometimes is given credit as the author. WorldCat often gives ascendancy to the organization which I think is wrong, and probably what was not intended. I agree that "clean and simple" is best. Which is why I don't like the Harvard, MLA, or Turabian style's of citation and reference, favored by this site. Q 12:57, 8 June 2008 (EDT)

ESM's Evidence! book has three relevant examples:

• Book authored by an agency (e.g., Bureau of the Census): list the name of the agency as the author.
• DAR Genealogical Records Committee report (local/state copy - unpublished): list a human being as the author.
• DAR Genealogical Records Committee report (national DAR library copy - published as a serial): no author.

Given this, I'd suggest listing a human author if someone's name appears somewhere on the front cover or title page. I don't think I'd suggest looking for authors in the appendix simply because most people won't go to that much effort. And if no human's name appears on the front cover or title page, to avoid possible duplicates (two different sources having the same Source page title) I'd suggest using the organization's (hopefully not too long) name as the author.--Dallan 10:33, 10 June 2008 (EDT)

Publication info:

Would this be "Essex Co. Quarterly Court Files, vol 4. leaf 63.", or perhaps the Genforum message Address?, or perhaps "Essex Antinquarian". (See note at end.)

Here I recommend placing the name of the message board, Scudder Family forum, message #, date accessed, and general URL for Genforum. --Beth 10:18, 21 March 2008 (EDT)

Call number:

   Is this applicable?  If not, should it be left blank?  or perhaps N/A should be inserted

This is also going to be moved down to the repository section, as it is the location of the source in a physical library.--Dallan 16:43, 22 March 2008 (EDT)

Type:

Is this "Will", "Bulletin Board Message", or perhaps something like "Court Will and Probate Record"

Types have been discussed several times by the source committee. Perhaps one should decide the purpose of the designated field type on WeRelate. If we know the purpose; then one can better define the parameters of the field. i prefer types that identify the type of record, deed or land, property, wills or probate or estate records, vitals (birth, marriage, death) etc. --Beth 10:18, 21 March 2008 (EDT)

"Type" here comes from the GEDCOM specification, where it is used to store what we're calling "format": book, article, electronic, film, fiche, etc. Since a single source could be in different formats in different repositories, this field either needs to be moved down into the "Repository" section, where each repository could have a different format, or we need to not track it at all. Thoughts?--Dallan 16:43, 22 March 2008 (EDT)

AHa! A different interpretation altogether. Since that's what was meant, perhaps "Format" would be better than "Type" Q 11:29, 23 March 2008 (EDT)

Yes, if we keep it I'll rename it "Format". I'm not sure how much tracking this information really adds though; I lean toward removing it.--Dallan 11:22, 25 March 2008 (EDT)

Repository name: [26 March 2008]

   Is this where the original record is kept (Essex Count Court House, I presume), or perhaps its "Essex Antiquarian", or perhaps "GenFourm"

Internet. --Beth 10:18, 21 March 2008 (EDT)

I notice that you've changed your mind on that. Originally you had "Genforum". That perhaps underlines the problem. The answer is not obvious, at least to me.

This question was addressed by ESM on a mailing list. My recollection is that the internet is comparable to the Allen County Public Library and Ancestry, Rootsweb, Genforum are sections in the library. I answered first "GenForum" because in my genie program; I choose to enter the "sections" because it enhances my search abilities in my program. So I believe we need to universally use "the library" or the "sections"; just need to decide as a group. --Beth 10:27, 22 March 2008 (EDT)

Somewhere in this discussion, I think I saw a note by Dallan to the effect that "Repository" meant a "physical repository"; somewhere that hard copies, or microfilm images, etc. were stored, but not an electronic medium like the internet. Perhaps there's a distinction to be made between "Physical Repository", and "Electronic Archives" Q 11:29, 23 March 2008 (EDT)

I think we'll have to make some exeptions to the general rule that "Repository" means a physical repository. Creating a single repository called "Internet" and asking people to put the homepage URL of the source in the "Location within Repository" field greatly reduces the number of non-physical repositories. That would be my preference. I could also see creating two Repositories: "Internet - free" and "Internet - fee", which is how GenSmarts divides things up so you can search for all online sources or just those in free sites.--Dallan 11:22, 25 March 2008 (EDT)

I really dislike the idea of an "internet" repository. That's completely useless. We talked before about when a website is a repository -- i.e. when it's like Jane Devlin's New England records site and holds transcriptions of hundreds of vital records. If we reduce repository to "internet" we remove the ability to collect information on a site like that. Same thing for Ancestry, NEHGS, and dozens of other very useful sites that should have repository pages, but are not themselves sources as they have been defined here. The way research is done these days non-physical repositories are way more useful than physical ones, so I don't understand the preference for reducing them.--Amelia 11:33, 25 March 2008 (EDT)

I'm open to this -- I hadn't thought that people would want to collect information about a specific website on a "Repository" page describing that website. What kind of information would you want to collect?--Dallan 11:55, 26 March 2008 (EDT)

I agree with Amelia no this. If what is meant here is simply calling the entire "world wide web/internet" a repository, then we defeat our purpose. A repository is, I think, a place you can go and expect to be able to retrieve specific things. In a conventional repository (say the Draper MSC) you can go to the physical location of the repository and ask to see the actual documents. You can do that today, and you can do that tomorrow, and you can do that for an indefinite period into the future. You can't really do that with the "Internet". It's there to day, and it may be there tomorrow, but there's little reason to think ANYTHING on the world wide web will be available to you for an indefinite period into the future. Hence, while it can serve as an information source, it can't really be described as a "repository". Some locations, like Ancestry, and perhaps "WeRelate" have more potential for persistence into the indefinite future than others. Specific sites such as that, especially if supported by the federal government, can be expected to be available forretrieval indefinitely. We MIGHT want to consider specific sites like that as repositories, but not the "internet" in general.

In the same vein, citing the "internet" as a source would not be a recommended approach to source citation. It has about the same utility as the ever populat "Various County Records". That is not a useful source citation since no one can follow it up to see the information itself. (If you haven't told people specifically where to look for the information, its not useful as a citation.) A citation has to be specific enough that others can follow it up and see for themselves. Q 12:45, 26 March 2008 (EDT)

Q says he agrees but for different reasons I think... My point was that there are any number of websites out there that are more properly called repositories than sources - most of Ancestry, most of Rootsweb, familyhistory.org, Google books, etc. -- they have very little original content (except perhaps trees), but hold book and record images or transcriptions. Ignoring their existence because they might disappear one day is silly. I think it's useful to have repository pages for these sites 1) because they have usage tips all their own; 2) because they are gorillas in the current genealogy world that should have a page where their strengths and weaknesses are discussed; 3) because if they do move or become defunct (a possibility I'm much less afraid of than Q), I'd bet a bunch of their data migrates somewhere, and the page would be the ideal place to deal with that. As far as what information to collect, the usefulness is in the usage tips, which are free form. But beyond that, URL and author or sponsoring organization are most important I suppose. --Amelia 19:19, 26 March 2008 (EDT)

We'll have a place on the Source page where you can list the repositories in which the source can be found. On each line, you'll be able to list the title of the Repository page along with the source's "location within the repository". The location field would contain a call number for offline repositories, or a URL for online repositories. So here we're discussing whether the repository title for all online repositories should be "Internet" (with the URL of the source listed in the "location within the repository" field), or whether we should create separate Repository pages for different websites. I can see creating different repository pages for different websites when you want to record usage notes about the website. I don't think we'd want to create separate Repository pages for all 50,000 websites with genealogical information though, because it would clutter up the Repository namespace quite a bit. How about encouraging people to create Repository pages for the larger genealogy websites but not for small personal websites? That would avoid the clutter.--Dallan 13:55, 27 March 2008 (EDT)

Repository addr:

Depends on the answer to the above, and could be any of several different things.

I would cite the primary source as an addendum to your source; but I do not recommend backwards sourcing. Your question also raises flags about WeRelate being too restrictive on dictating which category a user must use for certain types of sources. I do believe that Dallan should delete all of the message board sources that do not have citations and have them added individually. --Beth 10:18, 21 March 2008 (EDT)

I would solve this problem by putting the Will on Thomas Scudder's WeRelate page, and linking to that from any other people that need to cite this will. The reason is that, for the reasons you note above, a will does not fit well into the Source concept. Not only do the typical fields not apply, but this is a really specific source that applies to maybe a handful of people, and thus I don't think it belongs in the Source space at all. (If the digital library is going to encourage such things, that sounds unsustainable in the long run.) As a MySource page, I think it fits under the "miscellaneous" type where you do the best you can with the fields, rather than attempting to make sure the field definitions apply to this situation. I think the most crucial thing to include for WeRelate purposes is the Essex Court files, with a list in the text notes as to the places it has been reprinted. The fact that you found it at genforum is secondary to where the original can be found as far as others are concerned.--Amelia 23:19, 21 March 2008 (EDT)

Interesting, thoughtful response. Its also speaks directly to the underlying issue for this thread. See response below:

I think Amelia has hit on the crux of the problem. It seems that the "Source" for your will should be the Essex Court files, not the individual will. When we find a transcription of a census on Ancestry.com, we don't say that the source is Ancestry.com, we say that the source is the census, and Ancestry.com is one of the repositories. I think this case is similar: you found a transcription of one of the wills in the Essex Court files on a GenForum message board. So GenForum is the repository of a transcription of one of the items in that source. Since GenForum is probably a repository for only a few of the items in the Essex Court files, I wouldn't list it as a repository on the Essex Court files Source page. Instead, I would describe the transcription (author, URL, message date, etc.) in the description field of the digital library item you're uploading.

Alternatively, I suppose you could say that the source is the Genforum message board itself. By uploading the actual text that you're citing into the digital library, it makes citations of this particular item less "ephemeral", and so might justify citing something found in an online message board. But it seems to me that making the Source the Essex Court files is better so long as you state where you found the transcription.--Dallan 16:43, 22 March 2008 (EDT)

Some Other Discussions

Note: On citations, I'm not overly hung up about citing exactly where I saw something, as opposed to where the original is located. That is, Ultimately everyone should be using primary sources, but that's not going to happen. Most of us are mining the web for information, so by definition we are not looking at primary sources. However, I have no problem with citing the primary source (in this case the appropriate Essex County court records I think), as long as I insert something like "fide GenForum...." and cite the person who provided the information/message number/url. (And if its a truely ephemeral source with no hint of the primary sources, I probably wouldn't even bother with that. No point to that, as its not likely to be around very long for anyone else to look at.) Q 11:39, 20 March 2008 (EDT)

The answer to what should go into each of those fields, is quite dependent on what you think is the "Source" for this will. All revolve around what is considered to be a "Source":

• a) the item itself (the original handwritten will),
• b) the "repository" in which the original will was placed,
• c) the repository that houses the original will,
• d) the microfilm record of the will,
• e) the repository that houses the microfilm record,
• f) the Essex Antiquarian where a copy is said to have been published.
• g) the MessageBoard posting where I actually got the copy of the will
• h) the MessageBoard itself, that acts as a repository of messages dealing with "Scudder"?
• i) the person who left me that message, and recovered the information (from somewhere not stated) in the first place.

Any of these could be construed as the "Source". Perhaps all of them are, depending on purpose.

I'd say that B, C, and E should be "Repositories". I'll be adding a Repository namespace soon. A, D, and G are all pretty specific - it would be better to create a Source page for the Essex Court files and add information specific to the particular will you're uploading into the digital library item description. F could be a source, or perhaps better the particular article in that journal that you're referring to. H could be considered a source especially since you are uploading a transcription from it into the digital library, so your transcription has a greater chance of being preserved long-term (although I think I'd prefer creating an Essex Court files Source page and then explaining that I got the transcription from the message board). I is an author.--Dallan 16:43, 22 March 2008 (EDT)

The will could be placed directly into the article dealing with Thomas Scudder. The downside of that is that then you'd have to place it in every other article that refers to the will as well (for example, each of the children; spouses, wives, perhaps grandchildren, etc). So its probably a good idea to have the will itself in a separate location, to which the reader can be referenced whenever it comes up. The question is "where"?

• MySource
• Source
• DigitalLibary

And if the answer is that the will belongs in the DigitalLibrary, then there's a requirement for a "Source" entry that this is pinned to. (Can't be MySource, as the DLib ignores such articles).

Here, we might get some guidance from what has already been entered into the DigitalLibrary.

I wouldn't do that necessarily. It's just getting started, and most entries are from a single user who is operating under a deadline to get a bunch of material posted by March 29th. We may have to go back and fix up a few things afterward.--Dallan 16:43, 22 March 2008 (EDT)

Here's a link for an item, one of 50 or so similar entries

Abbott, John B. Freedmen's Savings and Trust Record, Charleston, SC, 1870

This item is in the

Community: AfriQuest

Collection: LowCountry Africana

Source: Register of the Freedman's Savings and Trust

Title: Abbott, John B. Freedmen's Savings and Trust Record, Charleston, SC, 1870

Most of the items currently in the DigiLib are related to these registers (ie, Account entries for various persons, recorded in the Registers). Thus the items here are related to specific individuals. What that tells me is that so far, one of the applications of the DigiLib is to record documents related to specific individuals. I think there are other applications, where the documents preserved may not be quite so specific (for example, a marriage register). But certainly the intent is to provide a place to store individual specific records---and wills would be an example of those types of records.

So the Digital Library is intended to house things like wills. That may, or may not be a good decision per your discussion, but its currently being used in that way as part of the start up.

But then we get back to the basic problem. If the will goes into the Digital Library, what is the Source in WeRelate that it is to be pinned to? Whatever that source is, it has to be in the "Source" namespace, because the Digital Library demands that. So now, we comeback to the central question---what is the proper source to use?

When I entered the will for Thomas Scudder I created Source:Will of Thomas Scudder, 1657/58. That was a mess up on my part. Since the item I was entering was the self same will as I identified as the Source, I ended up entering a title with the same name as the source: Viz:

http://www.werelate.org/dlib/browse?type=browsekey&sort_by=2&order=ASC&rpp=20&starts_with=W

So, while I was using Source:Will of Thomas Scudder, 1657/58 as a source of information in the article on Thomas Scudder, it probably wasn't a "Source" in the sense that the Digital Library was using the term. Possibly what I should have entered as a Source was something Like "Scudder GenForum", or perhaps "Records of Essex County, Massachusetts",or even "Essex Antiquarian".

And that brings us back to the question of what is the best source to use here? I don't know. Its a very complex problem. But if the DigitalLibrary is going to work, we need to sort this out.---or at least come up with some temporary rules of thumb as to how to handle things like this. The particular problem has its own particular needs, but really this is only an example of a larger scale problem that needs solution.

Q 09:34, 22 March 2008 (EDT)

Ameilias View [22 March 2008]

You're right, and this same issue underlies a great deal of the discussion above. Based on my reading of the consensus (or at least the Dallan-approved theory), the source is 1) Essex County files; 2) the Essex Antiquarian article; and 3) the message board. Theoretically, things like wills and birth certificates are also themselves sources, but for the purposes of the wiki, my personal opinion is that it's not sustainable or desirable to have those be Source pages themselves. Imagine what happens if the Source space becomes filled with documents that relate to only one person or family each - search and category pages would be drowned in ultra-specific sources. Particularly if the digital library can house those kind of documents, that seems unnecessary. If the source the Digital Library requires doesn't have to be the same as the Digital Library document (which is what I was assuming above, but in retrospect there seems no reason for that), then I think that source should be the places where the will is reproduced. In your case, you could link the will library doc to the three (or more) places it can be found as the sources. If you have to choose one, choose the most primary, the Essex County files. (I suppose a purist would say you have to link it to the exact place you copied it from, the message board, but as a matter of usability and helping people find the information, I think all available sources are appropriate).

That I would agree with. My conclusion is, perhaps consistent with yours, that what is meant by "Source" in the context of WeRelate, is more like "repository". In this case the repository might be GenForum, Essex Antiquarian, or Essex County Files---(but then, you have a slightly different view below). Pre-digital Library, there was nowhere else to place specific documents but in "Source" or "MySource" namespace---and there were problems with that no matter which way you went. Placeing the document itself (transcribed or image format) in the Digital Library solves this. Of course, I personally have no intention of transcribing every issue of Essex Antiquarian (much less all of the Essex County records) into the digital library. Possibly someone might choose to do that, (they seem to be doing something like that in the AfriQuest community) but I think stocking the digital library is going to be, for the most part, a process of building up of individual records to meet individual needs---a scattered approach rather than a directed one. If that's the way its to be done, then there's probably a need to clearly define how repositories/Sources, and such are to be named. Otherwise, there will be a lot of wasted effort as people create repositories for the same information but under different names.Q 16:12, 22 March 2008 (EDT)

I would call the Essex Court files a source, but not a repository. Personally, I would call the Genforum message board a repository (or perhaps it is better to just have a single Internet repository), but I could see that others might want to call Genforum a source. A particular message should never be a source. And individual records, especially when they can be thought of as items in a collection, should not be sources. Books, articles, collections of government/church records are all sources.--Dallan 17:53, 22 March 2008 (EDT)

Repositories, on the other hand, would be the libraries where one can find the microfilms or copies of the Essex Antiquarian, and those aren't sources (they go in the new Repository space).--Amelia 11:45, 22 March 2008 (EDT)

By "New repositroy space" I assume you mean "Digital Library". Then in that interpretation what is a "Source"? If its not someplace to go to find a specific document, and its not the document itself, what is it? [If I've misunderstood, please adjust my vision for me.:)] Because the Digital Library is Keyed to a specific Source page, there needs to be a clear cut explanation of what is included in the concept of Source, and what is not. Ultimately, I think the definition of "Source" is a description of where you have to go to find a specific document. That might be a specific library collection (the traditional sense of "repository"), but usually that means a book, a journal, or perhaps in the case of civic records, a County Clerks office,(etc). To me, at this point, the distinction between "source" and "repository" seems rather blurry. Q 16:12, 22 March 2008 (EDT)

No, we're going to add a new "Repository" namespace. The repository namespace will contain information about places where sources are held (hours of operation, address, phone number, etc.). So if you're going to Michigan and want to do some genealogical research, you could ideally get a list of all repositories of genealogical research in Michigan, along with links to the sources each contained. The County Clerks office in Ann Arbor would be a Repository, and it would probably contain multiple Sources (and some of those Sources might also be found in other Repositories, like the county genealogical society, or the Internet).--Dallan 17:53, 22 March 2008 (EDT)

JRM's Example [25 March 2008]

I have some similar sources to the one discussed in this thread. I did this work before any of the recent source discussion and just sort of used my best judgement on how to formulate it. It's Source:Tuttle - Bible of Joseph Sherburne Tuttle.--Jrm03063 10:44, 21 March 2008 (EDT)

Fantastic page. I nominate your source page to use as an example in the tutorial. Too bad the images cannot be shared in the digital library. --Beth 11:13, 21 March 2008 (EDT)

This page shows one thing that we're just not going to be able to have in the digital library -- the ability to link from names in the transcription to specific Person and Family pages. I really like that!

This source brings up a question for me. From what we've said earlier on this page, it might be better to move the source contents on a separate article and limit the source page to just citation and usage information. I can see the benefit of doing that for medium-large sources, but for small sources like this there's a certain convenience to seeing the citation information and transcription all at once. What do people think?--Dallan 16:43, 22 March 2008 (EDT)

I guess I don't have a strong preference for whether this is all on a source page or whether it is split into a source page and associated article. Is the rationale to streamline searching?--Jrm03063 18:24, 22 March 2008 (EDT)

I don't think there's a problem here with consistency. I think the "Source" card should be designed to meet the specific needs. Tom Chatt's suggestion for format several months ago seems reasonable, but different sources have different needs for descriptive purposes. As Dallan suggests, one of those needs might be in the case of relatively small documents (like a will) to place it directly on the source card. On the otherhand, a link to an item in the Digital Library would probably work about as well. So, I guess in the final analysis I'd be in favor of letting the need dictate whether the item appears on the Source card, or just in the digital library be dictated by the specific needs of the user, and what they found convenient/useful. Probably small items on the source card (for convenience), and in the Digital Library (for consistency). Q 11:47, 23 March 2008 (EDT)

One reason to put things on Source or article pages is so that you can add links to people/family pages, which you can't easily do in the digital library. I'm ok with putting transcriptions on the source page if they're short enough to fit in a single page comfortably. The reason for putting long transcriptions in articles is simply to keep the page size manageable. If you have a really long transcription you might want to break it up into several pieces, create an article for each piece, and link to all of the articles from the Source page.

Having said this, I think most Sources will end up being for rather large items: books or record collections, so seeing a complete transcription on a Source page will probably be rare (your page is one of the exceptions). It will be more common for people to upload individual records to the digital library and reference the Source page from their digital library item. Eventually when you bring up a Source page you'll be able to get a list of the digital library items that refer to that source.--Dallan 11:22, 25 March 2008 (EDT)

Repository name and Location within repository fields in Digital Library items [22 March 2008]

The above discussion raises an interesting point. We said earlier that we shouldn't include "Repository name" and "Location within Repository" in source citations on Person/Family pages because different people working on the Person/Family page might have looked up the source in different repositories. But for a digital library item there would be a single repository from which the uploaded item was obtained. Perhaps I should add "Repository name" and "Location within Repository" as fields on digital library items. In that case the Source for the Thomas Scudder Will digital library item could be Essex Court files, the Repository could be Internet, and the "Location within Repository" could be the URL of the message from which this transcription was obtained.--Dallan 16:43, 22 March 2008 (EDT)

More on Source format [21 September 2017]

Dallan has identified a number of specific document types that are routinely used by genealogists on this site and elsewhere. I've adapted his table by adding specific examples in the format he suggested, as well as at least one alternative format that might be used. I'm inserting a subsection for each of the document types to serve as a point for discussing each, as needed. (Note: I've converted this table into the traditional HTML format. I've never gotten comfortable with the Wiki Table style.) Also, the alternatives suggested are not necessarily meant as being better than the originals shown by Dallan; they are intended primarily to stimulate discussion; some of them may have certain advantages, some not. In evaluating the following, here are some criteria that I think define a "good source citation":

a. The citation provides enough information that the source can be located by anyone who wishes to. (No references so obscure that others can not find it.)

b. The source that is cited is invariant. (It will contain the same information today as it will tomorrow.)

c. The source itself is publicly available for an indefinite period into the future. (Sources that can't be seen by others are not useful. Private and unverifiable information is not useful. Citing a document that only you have ever seen, and no one else can see in the future, is not useful.)

As a general comment, I think we should keep in mind that the goal here is not to define a "good source citation" in the sense of what would be proper if one were publishing one's research. Rather, it's to come up with rules of thumb for naming source pages and collecting information in a way that is usable in the WeRelate system. Because part of that system includes (eventually) the ability to download identifiable citations, principles of "proper" citation have a role, but they are not the end goal. While we can have help pages or other guidance that tells people how best to fill out various fields, or what sources are best to cite, the system also needs to be as intuitive and usable as possible, or no one but us diehards will use it. To that end, I suggest the following other concerns are important in choosing naming conventions (and inform Dallan's proposals):

• Identifiability. In other words, you can look at the page title and understand what that it refers to (a book about the family or a vital record or whatnot). This is both because the page title is used in searches and category pages, and because a user looking at a page should be able to understand what the sources are without clicking through to another page.
• Uniqueness. As much as possible, naming rules should generate unique page titles, to avoid both technical and usability issues (but, as with person pages, this rule might be sacrificed in favor of the next).
• Predictability. If two people have the same source, the rules should be clear enough that they will attempt to cite it the same way. Thus, when one goes to type in the title in the source box and it shows the matching sources, each will know what to type to see if the source is already there.

--Amelia 21:06, 26 March 2008 (EDT)

I think there is a fundamental difference between our proposed definition of Source pages and the way sources are used in most GEDCOM files. In the proposed definition, Source pages are similar to cards in a library or archive catalog (e.g. the Family History Library Catalog). Source pages are generally created for collections of individual items: a collection of government/church records, all issues of a newspaper, or possibly multiple editions of a book with the same title and author. Details about a particular record found in the collection, a particular obituary in the newspaper, etc., are placed in the source citation on the Person or Family page or on the digital library item for that particular record. This allows Source pages to be re-usable by people working on completely different lines, and allows "source discovery" -- people can search the Source wiki for a place where their ancestor lived to find all sources that apply to that place. It will take some getting used to because in desktop genealogy programs we're used to including both source information and citation detail in the source record, but I think that separating large-grain source information from fine-grain citation detail is a very good idea.--Dallan 13:55, 27 March 2008 (EDT)

Basically, what you are saying is that you want to reserve the "Source" namespace for citations to collections of information, with the elements of that collection appearing in the digital library. That seems like a reasonable approach. One of the advantages is that it makes clear that the page articles in the source name space are bibliographic identifiers, not the actual item, or a copy of the items content. Those things would be found, instead, in the digital library. Q 08:54, 28 March 2008 (EDT)

When I first came to WeRelate, this is exactly how I used the "Source" namespace, as a bibliographic reference. I didn't see them as anything more, which has caused some misunderstanding on my part when I discuss this issue with others. See Manwaring as an example of what I add to a source page. If I had contents from the reference I wanted to cite, I would add them to the person or family page, not to the source page itself. --Ronni 11:58, 28 March 2008 (EDT)

Getting people to recognize the distinction between say a letter, and the book, email, message board posting, etc in which it was found, may be challenging. All are commonly perceived as a "source". Also, while some books do contain collections of discrete items, (e.g, Source:Thwaites and Kellogg, 1905, which contains many specific letters which might be individually cited), other books, can not be so easily categorized as collections (except as in the broad sense that they contain more than one thought.) Superimposing this concept on the "community" and "Collection" concepts embedded in the design of the Digital Library, may also create confusion. For example, what is the distinction between a "Source" which is a collection of things that might be cited, and a "Collection" as its used in the Digital Library? The same word is (or might be) used in both cases, but has very different meanings and implications. Q 08:54, 28 March 2008 (EDT)

Perhaps the way to handle this is to emphasize that "Source" namespace is used to identify specific documents, rather than the information they contain. That is, to emphasize them as "Bibliographic References". That would allow us to say "This is where I got this information", and then use the digitial Library to display the specific item of interest contained in the "Source"---be that a letter, a text passage, or whatever. Then it wouldn't matter whether something were a "Collection" or not. It would be whatever it was, and the main requirement would be that it provide enough bibliographic information so that others could locate that same source. Q 08:54, 28 March 2008 (EDT)

Exactly! And if the community doesn't understand the format that's been set up, then "wizards" will guide them. If a user sets up a new "bibliographic reference" (i.e., source page), perhaps the final step would be to ask if they have content they want to include in which case they would be directed to the DL. Perhaps "bibliographic references" (i.e., source pages) already in existence could have a "button" that directs users to the DL to add content. If they don't want to include it in the DL for whatever reason, then the content is instead placed on a MySource page. All of this done through a "wizard" to help maintain consistency as best we can. The "Source" namespace identifier should probably be changed to something else then to get us completely away from thinking of those pages as sources. --Ronni 11:58, 28 March 2008 (EDT)

The "Source" namespace identifier should probably be changed to something else. Ronni. I don't know if its possible to do that at this stage or not, but yes! that would substantially reduce confusion. Possibly something like "References"? Q 12:28, 28 March 2008 (EDT)

I like the Source = Bibliographic Reference idea. And I also like the idea of adding a button to the Source page that takes you to the digital library to upload an item for that Source. The Source namespace could be renamed to something else, but it would require a fair amount of effort. Let's keep this idea in mind and re-visit it over the summer once the new Source and digital library functionality have been in place for awhile.--Dallan 01:55, 30 March 2008 (EDT)

Perhaps the simplest solution would be to keep the term "Source", but make sure the explanation emphasizes that its meant to be a bibliographic citation, and that the actual item (or portion thereof) might be placed in the digital library. Q 10:05, 30 March 2008 (EDT)

Resource Specific Discussions

Books [27 March 2008]

The format suggested by Dallan looks like its based on the standard "footnote" format. This is a common style, and found in many of the works that are likely to be referenced by Genealogists. Traditionally, references like these are placed at the bottom of each physical page, and provided complete information about where the information was coming from, and often provided some additional commentary. This format is embedded in the data requirements for each article created in the person or family names space, and probably others. In that location it probably works as well as any.

However, for those creating more complex articles, where references to sources are embodied in the text of the report, this approach to source titles is a bit awkward. The main problem with it is that using the title of the source in the article title creates a very long string of text, the exact wording of which must be remembered and correctly typed, each time the source is called out. This can produce a very awkward looking display.

The alternative Author-Date style has an advantage in that it is fairly compact, but still provides a reasonable mnemonic cue as to where the information is coming from. In addition, it is much more compact, and hence more easily remembered. This style is nearly universally accepted in technical writing, and is, I believe, commonplace in genealogical writing. Q 15:53, 26 March 2008 (EDT)

I can see where you're coming from, but I think there is an inherent drawback to the Author, Date alternative. You would be surprised how many authors publish more than one book in the same year. Thinking about one local genealogical society I belong to, I can think of one woman who authored 4 books for us in one year. All of them would have the same Author, Date. Similarly, there are the "corporate authored" books -- those which have the genealogical society itself as the author. It is not uncommon for a society to publish more than one such book in a year.

That's always going to be a problem no matter what format you choose. I know of some works (same author, same Title) where there are many different versions published in different years. Each containing different information. It does you no good, for example, to cite the same author:Title combination, because you won't know which specific version the information was drawn from. In any case, the routine solution to that is to use a) a letter indicating the specific work, and/or b) a more specific publication date. Q 16:56, 26 March 2008 (EDT)

I agree that there are always going to be problems and exceptions with any system. However, in the case of differing editions of a book, they are just that -- different editions and, therefore, different books. The way to distinguish them is to include either the year or the edition after the title. Smith, John. History of Suchandso County, 2nd ed. --Ajcrow 17:36, 26 March 2008 (EDT)

A second drawback to the Author, Date alternative are the books for which have incomplete publishing data. You see this regularly with self-published works. Someone may have compiled a family history, typed it up, and ran off a dozen or so copies -- and they didn't include the date they published. --Ajcrow 16:36, 26 March 2008 (EDT)

While that does happen, there are workarounds for that as well. For example you can usually figure out the approximate date of publication, and use that, perhas as in "c1972". Not nice, perhaps, but workable. Q 16:56, 26 March 2008 (EDT)

I have to disagree that one can "usually" figure out the approximate date of publication. Something that was originally typed could be from the 1930s to the 1980s (if not a greater time span). Barring knowing when the author died (and, therefore, being able to say "before 19xx"), or the book referencing an event has having "recently happened" or "happened X years ago," it can be very difficult to come up with a meaningful approximation of the publication date. --Ajcrow 17:36, 26 March 2008 (EDT)

Possibly that depends on how precise one wants to be. "Circa" covers a fair bit of ground. However, I'm not particular advocating one format over another. I have my preferences, but others have theirs as well.

Remember that these are page titles. They must be unique, or we have to add information to them to make them so (i.e. the person numbers after names). Using Author. Title. is almost always unique. It's also always obvious. As several people have pointed out above, Author. Date is neither. I also fail to see what problem caused by the title cite is being solved. These links are seen mainly in 1) source cites and 2) footnote references on person/family pages. Neither place is starved for space, and in both cases, having the title visible is far, far more useful to the user than the date. (Anyone want to try to distinguish Jacobus 1969 from Jacobus 1971??). And if someone does have such a preference for a format that conveys less information to the user, you can always use pipes and display only the shorter citation without changing the default --Amelia 19:39, 26 March 2008 (EDT)

It sounds like the problem stated at the outset is how to cite sources in running text using a short notation. I think Source page titles will generally have to include the source title (although perhaps not the entire title if it is very long) for the reasons listed above. They have to be unique, and if we have too many titles of the form "Scudder, 1899a" vs. "Scudder, 1899b" because there were two books written by different people named Scudder in 1899, it's going to be confusing. So let's try to solve the problem of shorter source references in running text directly. There are three ways to do this currently:

• One approach is using "references" (see Person:Ella Grey (1) for an example using references). The first reference to a source on the page would include the full Source page title: <ref name="Scudder, 1899">[[Source:Scudder, Moses L. Records of the First Church in Huntington, Long Island, 1723-1779]]</ref>. Later references would be just <ref name="Scudder, 1899"/>. At the end of the page you would put <references/> to list references in a footnote section.

• Another approach that can be used if you have a reference to the Source page in the "Source Citations" section of a Person or Family page is to use ^{[[#S1|Scudder, 1899]]}. This creates a link to Source #1 as a superscript with "Scudder, 1899" as the underlined text.

• A third approach is to create a MySource page, say MySource:Quolla6/Scudder, 1899 and in that page put "#redirect Source:Scudder, Moses L. Records of the First Church in Huntington, Long Island, 1723-1779" in the text box. Now whenever you reference MySource:Quolla6/Scudder, 1899 you'll be redirected to the Source page.

I don't think any of these approaches is well-documented right now; there's still a lot that could be done on documentation. If none of these approaches works for some reason, perhaps we could come up with something similar along these lines.--Dallan 13:55, 27 March 2008 (EDT)

Journal/Magazine Articles [26 March 2008]

Same as commentary on "Books". Q 15:53, 26 March 2008 (EDT)

The problem of the same author having multiple articles in one year is going to be even more than having authors with multiple books in one year. At random, I pulled the 2004 Ohio Civil War Genealogy Journal off my shelf and found that Susan Lee has 4 articles (not counting her editor's messages) and Dan Reigle has 9. All of them would have the same Author, Date. --Ajcrow 16:49, 26 March 2008 (EDT)

This issue comes up in the technical literature with great frequency, especially in certain fields, where information is often disseminated in short notes. The answer that's normally used is to cite the works as "Lee, 2004a", "Lee, 2004b", etc.

That's a workaround when you're dealing with an established bibliography. However, to have that be the standard for articles here, if one were to have something from the third issue of a periodical, one would need to look at the first two issues to see if that person had authored something earlier. On top of that, how does one count a person's articles? In the case of Susan Lee, would one include her Editor's Messages or just her "articles"? --Ajcrow 17:28, 26 March 2008 (EDT)

Letters and Personal Journals [27 March 2008]

Letters are a bit tricky. While I have many hundreds of letters dealing with family matters, its the relatively rare instance when I would have the occassion to actually site one of those letters. Usually, the correspondance simply details research notes made by one family member or another---facts which can be confirmed with other sources which have a more public distribution, and so are probably better for purposes of documenting facts in many cases. However, I do make use of historical letters not in my personal possession. Here its common to have these letters in a repository or some description, and/or quoted in some other useful source. The Draper Manuscript Collection, for example contains many thousands of letters from TransAllegheny pioneers. These can be consulted directly at the University of Wisconsin, or indirectly through various publications, such as Source:Thwaites and Kellog, 1905. This presents some problems with regard to identifying the source used for a letter.

If one were working from the original letter itself, then Dallans format would probably be close to being perfect. The only addition I think I'd make would be adding the date of the letter somewhere in the title. Another alternative would from giving the source the title in the form of "From Surname to Surname, Date". Quite possibly the simple "Source:Surname, Date" would do quite nicely. The fact that its a letter would not necessarily have to be pointed out in the title.

But what do you do when you get a copy of a letter from a book such as Source:Thwaites and Kellogg, 1905? One solution would be to include the letter (assuming no copyright issues) itself in the Digital library, indicating both its immediate source (Source:Thwaites and Kellogg, 1905:page number) as well as its ultimate source (Source:Draper Manuscript Collection:QQNN12, for example). In the Digital Library, the source title would probably be something like Dallan's suggstion (Source:Smith, Daniel. Letter to William Preston) or one of the alternatives. (Smith, 1774 would work nicely since its short.) Then the Digital Library text could include locations where the letter COULD be found. NOte that "Could be Found". If one of those places was the location where someone ACTUALLY found the letter (a book, email message, website) happened to be included, then it wouldn't matter whether the location was generally accessible or not, as long as the ultimate location (Draper Manuscript Collection) was noted. This is a fairly elegant way to get around the problems between citing where you actually found something, and where others can go to view it. Q 16:50, 26 March 2008 (EDT)

I'd rather see a single Source page for the Draper Manuscript Collection, and possibly another Source page for the Thwaites and Kellogg book, and have people include details about the particular letter they were citing in the source citation on the Person or Family page or in the digital library item. I'm hoping we can encourage people to create Source pages for collections of letters rather than for individual letters. Source pages for individual letters aren't going to be very re-usable.--Dallan 13:55, 27 March 2008 (EDT)

Government or Church Records [29 March 2008]

Government and Church records may well present different problems for sourcing, and may require different solutions, possibly multiple solutions. Dallan's suggestion probably works best for most instances that commonly arise, though I think that it may be a bit better to go a little deeper than shown here. Most governmental records can be readily cited using a locational format, though not all. Coming up with a one size fits all format will be challenging.

Church records may be citable in a locational format also, but many of those records are found in collections that are not so amendable---that is, they are composited from churches in many different locations. In some (relatively few) instances, author date works well (Scudder, 1899, for instance, is the only known source for the information it contains, the original records not surviving.

Using Author Date here is a far, far worse idea even than using it above. It conveys no useful information whatsoever. The user would have to click to see what's a vital record and what's not, which is annoying at the very least. It's next to impossible to determine the "author" of most vital records collections to even use for the cite. And since no one is likely to know who the author is, it would be impossible to find the cite with the search functions and apply it. The reason for using the long form starting with the place is precisely to deal with those issues -- it makes the cite findable and clear as to what it contains.--Amelia 19:46, 26 March 2008 (EDT)

I'm confused why the location wouldn't work for both government and church records. Presumably, if you have the church record, you know where the church was (at least the state if not the county). Even if it is "composited from churches in many different locations," it is still the record of that one church. I'm also confused about why Author, Date works when the original record is not surviving. What are you citing if the record no longer survives? (One could also argue that even if the original was destroyed after you viewed it, it doesn't change that as your source. You just can't look at it again.) --Ajcrow 20:26, 26 March 2008 (EDT)

In many cases it will work. But there are a good many examples where it would not. As an example, the Presbyterian Historical Society in Philadelphia includes as many of the early records of Presbyterian Churches as possible. They brought them together in a single repository which offers the researcher great convenience. Unfortunately, not all of those church records survive. In many cases, what exists today is not the records for the churches themselves, but from the Presbytery to which they belonged. Thsu the same document will contain references to a number of specific churches (including for example, lists of elders). Because those lists are from different churches, different counties, different states, identifying those sources locationally, is highly misleading. As I say, locational names for church records would work in most cases, but not in all.

That is true. However, I think you need to plan for the most common examples and figure out how to handle the exceptions, which could be on a case-by-case basis. In the case you give above, I could easily see handling it not as a church record, because it doesn't belong to a particular church, but as a manuscript collection. (Which we should probably have a format for.) --Ajcrow 21:09, 26 March 2008 (EDT)

I think manuscripts would fall into the same type as "Letter/Journal", which should be though of as the type for unpublished works. I labeled this type with a couple of examples, but perhaps that doesn't convey the point well. How can we rename the "Letter/Journal" type to reflect this general notion? Is "Unpublished works" too broad? "Manuscripts/Documents"? "Unpublished Manuscripts"?--Dallan 13:55, 27 March 2008 (EDT)

I don't think that "Letter/Journal" intuitively points a user to use it for manuscript materials. Would "Manuscript Collections/Unpublished Works" be too long? My thought is that by including the word "collection," it would be a bit clearer that you're looking for sources to the collection as a whole, not a specific letter within a collection. --Ajcrow 17:34, 27 March 2008 (EDT)

That's getting pretty long; something shorter would be better if it still gets the point across. If not, we could certainly use "Manuscript Collection/Unpublished work". (I think I'd make it singular since collection implies a set of documents already.)--Dallan 01:55, 30 March 2008 (EDT)

But of course, exactly the same argument probably works for Government records as well. Some Government records are also compendiums from different locations. Coming up with a one size fits all solution is going to be challenging in either case. In most cases "Locational" fits most circumstances I suspect, but it won't fit all---and the exceptions are probably significant. Q 20:39, 26 March 2008 (EDT)

As far as I know, the record collections usually follow some jurisdictional hierarchy, although the hierarchy may not be the same as the political hierarchy especially if they are not government records. You've already pointed out that Church hierarchies often differ from the political hierarchy (parish, presbytery, etc.). I'm told that military records also are sometimes kept along different hierarchies. One possible way to address this is to create Place pages for those hierarchies, which User:TomChatt has done for Place:Scotland (this page includes a nice graphic outlining the relationship between parishes, presbyteries, towns, etc. and is worth looking at). Parishes are located in counties (primary jurisdiction) as well as presbyteries (alternate jurisdiction). Presbyteries are located within synod's. Scottish church sources might then be titled using the parish or presbytery "place" that they cover. This approach has the downside that sources at the presbytery level aren't as easy to find as sources at the parish level because the titles of presbytery-level sources wouldn't include a county, but it has the benefit that the jurisdictional hierarchy for church records is clearly brought out, and understanding this hierarchy has got to be useful for people doing Scottish research. It seems using location works at least as well as using author, because in those cases where records don't follow political jurisdiction boundaries, we can represent in the title the jurisdiction that they do cover.--Dallan 13:55, 27 March 2008 (EDT)

Newspaper Articles [27 March 2008]

Most Newspapers are highly location specific (USA Today being a notable exception), so its natural to use a locationally driven format for the title. The simple format shown in Dallan's example has much to commend it, with very little added value from the alternative approach shown. Dallan's approach would work well within the context of the Digital Library where you have to pin a specific item to a separate source. Thus, if a newspaper article were the item of interest, included in the Digital Library, It would be better to pin it to something fairly simple following Dallan's Format. So you would have "Source:United States, North Carolina, Burlington. Burlington Times News" identified as the source, and then the subject identified as "7 November 2007, Obituary for Clara Sentell Hutchenson" or some variation such as "Obituary for Clara Sentell Hutchenson, 7 November 2007". ( You probably could not do t his particular example in the Digital Library since that would create some copyright issues.) Q 20:12, 26 March 2008 (EDT)

In the digital library you have the Source title and also the Item title. I'd enter the source as you have - one Source page for all issues of the Burlington Times News, and then enter the title of the digital item as "Hutchenson, Clara Sentell, Obituary, 7 November 2007" or something similar. That way people browsing the digital library by Item title (just added this feature yesterday) will be able to jump to "Hutchenson, Clara Sentell" in the browse list to see all items for people named Clara Sentell Hutchenson.--Dallan 13:55, 27 March 2008 (EDT)

Message Boards and Web Sites [27 March 2008]

This type is going away; we'll put these under "Miscellaneous" when they correspond to original sources and not online repositories of offline sources.--Dallan 14:11, 27 March 2008 (EDT)

Family Trees [29 March 2008]

By "Family Tree" I'm assuming that what is meant is one of the tree's found on a site like Ancestry, GenCircles, personal websites, and the like. Whether or not the site itself persists through time, the nature of such trees is that they are variable. They undergo more or less continuous change, and I don't think you can expect to revisit them to verify that they show a particular piece of information. They may, or may not. A family tree published in a hard copy format, on the other hand may have more lasting value as a source, particularly if it documents the basis for facts shown in the tree. Sometimes this is the case, and its not uncommon to accept such "trees" as authoritative. (Whether that's wise is another matter.)

While I use web based tree's routinely, its mostly to get a feeling for what others have found, rather than to use them as a definitive source. Its rare for me to cite them as a source of information per se. If I do, its because I think I particularly need to acknowledge someone elses work, but I wouldn't expect anyone to revisit that "source" with the idea of using it to verify or validate my own work. Rather, I'd expect them to look at the other sources I provide for that purpose. In most cases, especially family trees posted on a web site, the exact tree that was used at a specific point in time probably can not be revisited at a later date, because a) the tree won't be online, or b) if its still online, it will have changed. Q 20:27, 26 March 2008 (EDT)

Should there be a separate category for databases? One might say that it's just a variant of a website, but it's not always accessed thru a website, and it may be accessed thru more than one website. The workhorse that comes first to mind is the Social Security Death Index. It is maintained and published by the SSA, and I think several websites provide a query interface to it. It is ephemeral, in that it is constantly being added to and occasionally corrected. I think there are a number of index databases that exist and are of use to genealogists. To think of a few more off the top of my head, there's the FreeBMD project in the UK, as well as a free census project. There's an Italian genealogical society in New York that's actively building a database of New York marriages. What would be the appropriate way to make a Source page for these such things, and what would be the right way to cite them? --TomChatt 04:36, 27 March 2008 (EDT)

I've added a "Miscellaneous" source type, which will allow for these things. Social Security Death Index would be a good "Miscellaneous" source to add. Other things like the New York marriages database though might be better off listed as an online Repository for Source pages for the original record sets, for example "United States, New York, Albany. Marriages" -- assuming marriage records in New York are kept at the county level. In that case the website would be listed on each of the New York county marriage record sources as one of the Repositories where you could find those marriage records.--Dallan 13:55, 27 March 2008 (EDT)

I think I've been the main person arguing that we need a source type for "Family Trees", mainly to support uploading family trees to the digital library because we currently have a requirement that every item uploaded to the digital library must be linked to a Source page. I'm re-thinking that requirement. If we make the Source link on digital library items optional, it simplifies things. People would be able to upload individiual letters, photos, and their family tree to the digital library without creating Source pages for them. If we do this, I'd be ok with dropping "Family Tree" as an explicit source type. People could create MySource pages for family trees, which is probably more appropriate for most trees, or for well-documented trees appearing on stable websites, they could create Source pages using the "Miscellaneous" source type. What do others think?--Dallan 13:55, 27 March 2008 (EDT)

Optional "Source" page may be the the way to go with the Digital Library. It would simplify somethings. But if its optional, then you run the risk of folks simply not bothering to connect to a "Source" page, even when that makes perfect sense. (If the source page hasn't already been created then at this point in the process you'd have to drop what you were doing and create the source you were using. Since would be much easier to simply ignore the need for a source page, I suspect that's what some folks would do. There are, however, some other problems with this that I'll discuss on the Digital LIbrary testing page. Q 08:54, 29 March 2008 (EDT)

That's true. We can encourage people to change their behavior by making it as easy as possible to use good practices. But we can't require change. Hopefully most people will link to the correct sources if we make it easy enough, say by populating the Source wiki with the more commonly-used sources?--Dallan 01:55, 30 March 2008 (EDT)

Let's move to some actual examples [30 March 2008]

There has been some great discussion about what well done sources should look like. A lot of very good ideas. I'm concerned however, that we're starting to go overboard on this. Making it too hard to create a decent source citation seriously undermines our goals. I think it would also be a good idea to start shaking down some of the ideas so far discussed, in the context of a concrete example.

Savage [15 April 2008]

My first thought is Savage. It was already excellent due to the work of User:knarrows. My contribution is to add a section on usage customs and examples. What do folks think of this?--Jrm03063 14:46, 28 March 2008 (EDT)

I think that's an excellent example! --Ajcrow 15:38, 28 March 2008 (EDT)

True, at least I think so, but let's get specific:

• Is it properly named (Author then title)

• Is the name chosen too long, too little, or just right as a fragment of the author's formal title?

• Should the title portion of the name be in quotes?

• Is the structure of this content (citation, repositories, preface, usage customs and examples) a good general model for other sources?

• With specific regard to the idea of a usage customs and examples section -

 * Are the sorts of notes there useful, or should it simply be a couple of good examples?

 * Can someone think of a less unwieldy name?

My two cents: 1) There should be a period instead of a comma after James. 2) The title is a correct combination of the correct title of the book (i.e. it should not be shorter as that would do damage to the recognizability of this very well known title) while removing the excess part that's not usually used anyway ("Showing Three Generations of Those Who Came Before May, 1692, on the Basis of Farmer's Register"). 3) No. Book titles don't go in quotes, and that would just add more confusing characters. 4) I added the publication info into that field. This is also an odd case where Google says the proper citation includes three names, but the book itself has only Savage. Which means I think everything about this page is correct, but it's perhaps a confusing example on the treatment of multiple authors. 5-6) I added a more traditional "usage tips" section that comments on using the source itself. I think the "usage customs" section is a good idea (particularly for this source), but it's going to be overkill for many sources. I changed the section title to "citation examples" to distinguish it from the usage tips that show up by default everywhere else.--Amelia 10:53, 29 March 2008 (EDT)

1. I agree with Amelia that there should be a period after the author instead of a comma.
2. I agree that subtitles of books should be left off, but included in the "citation" portion
3. No, we should not use quotation marks in the titles for the reason Amelia gave. For this same reason, that's why I like a period instead of a comma after the author's name -- it separates the author from the title more clearly. (Savage is an early source page I created and I was following WeRelate title conventions at the time, hence the comma and not the period).
4. As to handling multiple authors, I prefer just one author's name in the title with the other authors listed in the citation portion.
5. I like the usage tips and citation example portions. Both are extremely helpful. --Ronni 11:16, 29 March 2008 (EDT)

Now that we've got all the bells and whistles we asked for, what do we do with them... Google Books, the FHC, and WorldCat should be set up as repositories. But what do we do with the online transcription, which is where most people are going to want to go first when they get to this page? USGenNet could be a repository, although it's a bit of a stretch. Do we in fact use the dummy "internet"? Leave it blank and just put in a URL?--Amelia 21:06, 11 April 2008 (EDT)

Good question! :-) I've been leaving the repository blank and just putting in the URL.--Dallan 17:29, 15 April 2008 (EDT)

NEHGR [4 April 2008]

Great, I've modified Savage and the examples noted as suggested. Let's move to a harder one. Let's take on the ever popular NEHGR. It has been properly noted as not a source in itself, but it needs a source page in order to support the various papers that appear within it. So I consider it a source that should never appear (alone that is) in a page source citation. Please kick the tires on it and let's keep moving! --jrm03063

This is terrific! I've been thinking along these same lines -- that it's time to implement the functionality we've talked about so we can get some experience using it and make further changes from there. Now that the digital library is good enough for the Lowcountry Africana launch, I can finally finish implementing the source wiki pages. They should be ready by the end of next week. Just one question from the current discussion: what's the consensus on multiple authors in the Source wiki page title? We've talked about:

• first author only (e.g., Source:Savage, James. A Genealogical Dictionary of the First Settlers of New England)
• up to two authors, with et al. added if there are three or more (e.g., Source:Savage, James, O. P. Dexter, et al. A Genealogical Dictionary of the First Settlers of New England)

What's the consensus?--Dallan 01:55, 30 March 2008 (EDT)

As to multiple authors, I would prefer the main (or first) author only listed in the title unless there's an advantage to adding additional authors in the title that I'm not aware of.

Regarding the New England Register as jrm's example, this is where I get confused. <g> I don't understand why The Register is not considered a source. I realize the Society is a repository and not a source as Amelia has noted in this "source page". But visiting the NEHGS site and viewing The Register, I notice they give citation information: The New England Historical and Genealogical Register. (Online database: NewEnglandAncestors.org, New England Historic Genealogical Society, 2001-2007), (Orig. Pub. New England Historic Genealogical Society, Boston, MA. The New England Historical and Genealogical Register, 158 vols., 1847-2004.)

If The Register is not a source, why is NEHGS giving us citation information on it? Unless you're saying that the magazine itself is not the source, but the actual article within the magazine that contains the information you are citing is instead? In any case, I would think of The Register as a Master Source and then in the citation details on my Person/Family page I would add the article title, author, page, etc. So, to address your example, I would simply make the following page: Source:The New England Historical and Genealogical Register (which already exists, but needs some tweaking). --Ronni 03:16, 30 March 2008 (EDT)

No disrespect intended to other contributing authors, but our primary string reference needs to be dominated by the source title, not the authors. We have the benefit of having a full backing source page so everyone who deserves credit can get it there. Using only the initial author is certainly the convention observed in my academic experience, so it seems fine to me.

Carrying on with sources, I'm afraid I bungled the link the NEHGR in my last posting here, so I'll try again. I tweaked it a bit yesterday to be more like what was done for Savage, adding some examples which I expect to be the main focus of discussion.

My thinking is that, rather than trying to get folks to know a lengthy set of rules (my eyes just glaze over when I start reading that sort of stuff), we flesh out important sources with some decent representative examples.--Jrm03063 09:28, 30 March 2008 (EDT)

I agree jrm. I only have two minor issues with the NEHGR page: I don't like links in the sidebar for aesthetic reasons (I would include them under the "Repository" heading somewhere). Also, the citation examples jrm gives are great, but they are not the only way to cite this source, or any magazine, journal, newspaper, etc. As I mentioned before, I doubt I would make a separate source page for every article I reference, in which case, I would link to NEHGR in my citation, but would also include the article and author in that citation. The examples are suggesting that a separate source page needs to be created. As to the title format and how the rest of the page is setup, things look good. --Ronni 13:39, 30 March 2008 (EDT)

I've heard strong opinions that NEHGR should never appear without narrowing the reference with author and title. I've sort of come to accept that opinion, from looking at stuff like "NEHGR, v5, p11-12" (sure, you can find it, but it doesn't exactly pop out at you). I've added another example (Person:Thomas Welles (3)) that follows the general form of the others, but doesn't create a new source page for the article (I just use the title only form but still insert NEHGR as a wiki link). Does this seem an acceptable alternative?--Jrm03063 16:39, 30 March 2008 (EDT)

Oh, I agree with you about NEHGR should not appear alone in a citation. My point was that the article doesn't have to have its own source page. This same thinking would apply to an article from a newspaper or obituaries even. While the obit would be the source, I again think of the newspaper as a *master source* but include in the citation details the title and author of the article. The newspaper would have a source page, but the individual obits and articles would not. Using the Digital Library, it would be easy then to create a Newspaper Collection adding individual obits and such.

Seeing your example though in the "title only" form, I wonder if an additional field can be brought in where we can add *article* titles and still use the "source" form? --Ronni 17:22, 30 March 2008 (EDT)

It sounds like the consensus is to put just the first author in the Source page title. I'll go with that unless someone disagrees.--Dallan 10:12, 31 March 2008 (EDT)

I agree with creating a single Source page for a newspaper and having all obits refer to that single source page. I'm less sure we want to do that for articles though, because articles have explicit authors and titles, whereas obits do not. I thought we discussed earlier in this page about creating separate source pages for each article (and we could create a "Master" source page for the periodical). If we don't want to create Source pages for individual articles, then I should remove the "Article" source type in the tables above.--Dallan 10:12, 31 March 2008 (EDT)

My preference is for a source for every article. Still, I think a useful source may not require a specific source page. The second example I added to the NEHGR source demonstrates how I think that could be done (title only for the article, with a wiki-link for the periodical as the first part of the "page" field.--Jrm03063 10:34, 31 March 2008 (EDT)

Go with the concensus Dallan on creating source pages for articles. Mine was just one differing opinion. <g> We do want to make source pages for articles and if creating a source page for every article I cite needs to be done, then I can do that too. --Ronni 11:27, 31 March 2008 (EDT)

Its likely that if there are multiple articles of interest in a compendium (of whatever description) its the individual articles that are going to be most commonly cited. However, there are going to be cases where you'll need to have a source to the compendium itself. I can't see much advantage of restricting things to one type or the other. If a Source card is prepared for a specific article, its still going to need to reference the compendium in which it was found, be it Journal, Symposium, or whatever. If its a Journal article, for example, you might want people to know where the Journal was published, it publication history, where copies are held, etc. You'd not want that to go on each source card for the articles in the Journal, but it would be appropriate (and useful) to have a separate card providing that for the Journal itself. Q 11:52, 31 March 2008 (EDT)

What Q said -- pages for both the journal and the articles separately have their uses. As for a citation format, in the interest of being predictable, consistent, and short, I vote for the general standard of Author. "Article Title", with a pin cite (page) added where available. I don't think NEHGR, the year or anything else needs to be added to the person/family page -- that's all on the source page, and the less people have to type, the more likely it's going to happen. Also, we don't put publication detail on book cites, and the journal information is similar in this situation. The Tibbetts cite above, for example, I don't think is a model because it generally cites three whole volumes of the journal. The lack of pin cite is my bad, but I don't think the volume cites make up for it.--Amelia 22:51, 31 March 2008 (EDT)

I'm all for the economy of avoiding redundant information. The problem I see is how to cite a specific page in a paper? You can certainly provide the page number, but that's not meaningful without a volume. And the volume is sort of strange without indicating that it came from NEHGR. A naked page number doesn't really cut it either, since articles can be spread out in different editions and, perhaps, different volumes. I admit it's a slight hassle to preface the volume and page reference with a periodical source reference, but I don't see how it can be avoided in a useful reference. I suppose there are programmatic ways to deal with it (if an article source page references a publication "master" source, the standard master source abbreviation is automatically dropped in for you), but that seems sort of heavy-handed and I hate to add software burdens...--Jrm03063 11:13, 1 April 2008 (EDT)

I was thinking we'd have two fields on the Source page for this: "Periodical/Series Name" field, which would contain the title of the Source page for the periodical, and a "Volume/Page number" field, which would contain the volume and optionally the page number. Would that work?--Dallan 12:09, 1 April 2008 (EDT)

Well, that's great for an article's source page. My concern is what the source looks like on a person/family page, where you want to cite some subset of the pages in the entire article. If that is the basis for something that gets tossed in after the article title and before the page stuff, I think that's effectively what I've been doing by hand w/NEHGR.--Jrm03063 12:20, 1 April 2008 (EDT)

I'm not quite sure I understand the comment. I think you're saying that you plan to enter the NEHGR reference in the Page field of the source citation on the Person/Family page so it is displayed on the Person/Family page (is this correct?). I think that's fine. But in general, not all of the information necessary for a complete citation will be part of the source citation on the Person/Family page. Fields like publisher, publication place, additional authors, and the periodical/series name will be on the Source page unless you copy them into the source citation fields on the Person/Family page.--Dallan 23:59, 1 April 2008 (EDT)

If we're going to strip this down to the absolute basics, such that we're not even going to have an indication of what periodical contains a referenced paper, then I suggest an absolute bare-bones approach:

    Doe, John. Genealogical Paper of the Century, 12:352-353

Or:

    Doe, John. Genealogical Paper of the Decade, vol 12, pp 352-353

While quoting the title of a paper is customary, that usually happens in the context of a citation where the name of the containing periodical also appears. If there's nothing to distinguish because we're relying on the backing source page, the quotes are just taking up space.

If this is what folks had in mind, I'll modify the examples and we can all take another peek...--Jrm03063 17:22, 3 April 2008 (EDT)

So the Source page title would be "Doe, John. Genealogical Paper of the Century" and the Volume/Page# field in the source citation on the Person/Family page would be "12:352-353" or "vol 12, pp 352-353". That makes sense to me.--Dallan 11:14, 4 April 2008 (EDT)

Merging source pages [2 April 2008]

I plan to merge these 2 source pages.

• [1]
• [2]

What does the number (267288) refer to in the second source page? --Beth 17:46, 1 April 2008 (EDT)

That looks like the right thing to do. Here's another source that could perhaps be merged as well: Source:Reprint of official register of land lottery of Georgia, 1827 (440216). As for the 267288, when I created the Source pages from the FHLC two years ago, I set the Source page title to the FHLC entry title. This resulted in a number of duplicate Source page titles, so I added the FHLC ID to the end of the duplicate titles to make them unique. In this case there were two sources titled "Reprint of official register of land lottery of Georgia, 1827", so I added 267288 to one and 440216 to the other. The correct title should be Source:Houston, Martha Lou. Reprint of official register of land lottery of Georgia, 1827.--Dallan 23:59, 1 April 2008 (EDT)

I have some questions; why is the author/compiler part of the source title? If the author/compiler is part of the source title does one enter information in the field - author? Is it necessary to enter the author's middle name? Are we to ignore the distinction between author and compiler?

What has been decided regarding capitalization in the source title? In your example only the first word and the place name are capitalized.

--Beth 19:21, 2 April 2008 (EDT)

The author is part of the Source page title to make the source page easier for others to find later -- the idea being that it's sometimes easier to determine a book's author than its exact title, and once we rename all of the Source pages to the new format, if you enter the author's name in a source field, you'll get a drop-down list of all books written by that author. Yes, we ignore the distinction between author and compiler, and I think it would be best if you entered the author's middle name if that's the way it appears on the book.

Sorry - I messed up the example. Yes, title words except for short prepositions should be capitalized. (I just finished writing the function to automatically capitalize titles yesterday.) So the Source page title should be: Source:Houston, Martha Lou. Reprint of Official Register of Land Lottery of Georgia, 1827.--Dallan 14:58, 3 April 2008 (EDT)

Okay, I successfully merged the three pages and renamed the title per your direction. There are lots more but have not dealt with them yet; they are county specific but still from the same source. Do I count this for volunteer time or just my usage of WeRelate? I investigated the source because of transcribed data from GeorgiaGenWeb from this source; but I will not use the source at this time because my data is not from the source, but rather a transcription from this source. --Beth 20:54, 2 April 2008 (EDT)

I think you could count it as volunteer time, thanks!--Dallan 14:58, 3 April 2008 (EDT)

State GenWeb Archives [4 April 2008]

There are separate source pages for each county in Georgia for the GenWeb Archives. I assume that this is the same for all of the state GenWeb Archives. Before I cite my source from the Walton County Georgia GenWeb archives; I suggest that we consider having one source page for each state. The user could add the county link to the state source page and include county specific information in the citation.

Here is the link to the Georgia GenWeb Archives [3]. Scroll down toward the bottom of the page; all of the counties may be accessed from this one page.

Opinions. --Beth 08:17, 3 April 2008 (EDT)

I looked at the Walton County GenWeb page and it appears that most of the pages are transcriptions from other sources, and the GenWeb page is actually a Repository for records from those sources. (I'll add the new Repository namespace tomorrow.) Some pages list the source where the transcription is from; for example, this page. If I were citing that page, I'd use the "Houston, Martha Lou. Reprint of Official..." source mentioned above and add the GenWeb page as a Repository to that Source page (although you'll have to wait until tomorrow to do this).

Hi Dallan, thanks for the information. I know that you and the rest of the team have reached a consensus opinion on this, which is fine but I don't plan to cite my data this way. This is a temporary citation (source) until I visit the Georgia archives to obtain an original source. I can either enter my citation under mySource or title only; whichever you prefer; but I refuse to cite a source that I have not examined.--Beth 17:01, 3 April 2008 (EDT)

Reading through the concensus opinions it seems that I can use MySource for GenWeb sources. I prefer that to title only; there is no "choose" feature with the title only.--Beth 18:17, 3 April 2008 (EDT)

See discussion on this topic below.--Dallan 11:14, 4 April 2008 (EDT)

It would seem that y'all have decided to remove web pages from sources. I created a source for a web page that is a cemetery association with a transcription of the cemetery and some photos of the tombstones. There is more than one family line buried in this cemetery; so there might be someone else that wishes to cite this cemetery, but probably not that many people. I can change this to a MYSource.--Beth 18:17, 3 April 2008 (EDT)

One of the main goals of the Source wiki is to help people discover sources that are available. If you change it to a MySource that's not going to help anyone, because MySources are going to be generally low-quality GEDCOM sources and most people won't search them (I'm going to separate searching Sources from MySources). How about renaming it to "Reversed place name. Cemetery Data" and adding the webpage as a repository of this information? Then anyone looking for sources related to the place would be able to find your source page.--Dallan 11:14, 4 April 2008 (EDT)

Also when you redo the sources what happens to links to sources that are deleted.--Beth 18:17, 3 April 2008 (EDT)

We'll redirect the old names to the new ones. I don't plan to delete Source pages that others have referenced.--Dallan 11:14, 4 April 2008 (EDT)

Other pages though don't list the source where the transcription is from. This raises a good question: What source do we use for record transcriptions on web pages that don't cite an off-line source? Any thoughts? Would it make sense to create a generic source page; e.g., Source:United States, Georgia, Walton. Vital Records, and add the GenWeb page as a Repository for that source, or is there a better way of handling this?--Dallan 14:58, 3 April 2008 (EDT)

The Rootsweb Archives are indeed a wonderful resource for genealogists. I use them frequently. However, I would be cautious about citing something in the GenWeb archives as a source. I might use it to get pointers to the original material, but I would cite the archives themselves as a source only where the material was well sourced, and the original not available to me. The reasons being:

• a) they often fail to provide anything like a pointer to a sound source,
• b) their content is not uniformly reliable (some very good, but I can point to some that are very poor transcriptions),
• c) they almost universally claim copyright over the material. (That they claim copyright is, I think, bizarre, since most of the information is in fact out of copyright. Yet the copyright banner is there, and if you transfer material whole cloth then you theoretically have a copyright issue. I suspect there's a lawsuit waiting to happen on that, but I think the concensus of opinion on this is that there's no explicit case law---so people can argue whatever they want.)
• d) Any links to these sites are subject to change without notice. There is no guarantee that they will even be housed in the same organization. Spending time fixing broken links is not something I'd personally want to do. So, as far as possible, I try to link to sources (such as Google Books, or the Million Books Project) where I think there's reasonable likelihood that the links will survive an indefinite time into the future. That's not always possible, and you will find the occassional reference to something in the GenWeb archives et alia in my articles.

While I think we should take appropriate advantage of the generosity of those who have placed things in the Rootsweb archives, I think the above factors need to be considered when using this resource as a "source". I would also be very cautious about transfering something out of the Archives whole cloth into the Digital Library.

Q 16:22, 3 April 2008 (EDT)

I never intended to enter any items from the Genweb site in the Digital Library. I planned to cite information from the transcribed record for people who were fortunate drawers in one District. --Beth 17:01, 3 April 2008 (EDT)

Citing Online Sources [5 April 2008]

This question has come up a couple of times already in the last couple of weeks: how do I cite a source that I found online, when I either don't know the corresponding offline source from which the online material was transcribed, or I haven't personally seen the offline source and I am not sure about the quality of the transcription? These are actually two separate questions. I know the ideal answer is that we not cite online sources because they are ephemeral and if they disappear there is no place for someone trying to repeat the research to verify the information. And getting people to come up with consistent Source page titles for online sources is problematic. For example, the title of the Walton County GA GenWeb page is "Walton County, Georgia," the title of the archives page is "Walton County Georgia USGenWeb Archives", and the title of the Deeds/Land page in the archive is "Walton County Georgia Deed & Land Records" - which do I use? And talk about ephemeral - these page titles could change easily. But unless and until I look up the information in a more permanent repository, it seems that I'd want to be able to leave a note to myself and others where I found the information.

Here's what I'd like to propose to address the issue:

• If I find information online that is a transcription of an offline source and I know the title of the offline source, I should create a Source page for the offline source. Then in the source citation on the Person/Family page I can enter the URL where I found the transcription in a new "Where found" field, which is just a text field. If I or others come along later and find the record in the offline source, the offline location also can be entered in the "Where found" field if desired. This covers the case of I'm the only one who has found this record, and I want to let others know exactly where I found it as well as the case of people have found this record in different places. The "where found" field could be used also to capture distinctions such as those expressed under the Purpose of Sources heading at the top of this page.

• Instead of adding a new "Where found" field to the source citation, we could rename the existing Description field to "Description / Where found" and use that field. That's less clear than a separate "Where found" field, but would be more compatible with exporting a GEDCOM. After thinking about it, I think this is my preference.

• Now suppose that I find information online that was transcribed from an offline source, but the title of the offline source isn't given. Source page titles describe what the source is, not where it is found. So until a better title is found, I enter a title that describes the source to the best of my ability. For example, if I find a record on the Walton county GAGenWeb vital records page (sorry to pick on this example), I would title the Source page according to the Government/Church records format: "United States, Georgia, Walton. Vital Records". I would enter the URL of the Walton county GAGenWeb vital records page as a repository on this Source page. And I would enter the URL in the "Where found" field of the source citation on the Person/Family page. Then others coming along later who find this Source page could add additional repositories for this source.

• As an alternative to the above, there's the MySource namespace. But I'd like to encourage people to create Source pages when possible so that others trying to discover what sources are available can find the Source pages. I'm going to separate searching Source pages from MySource pages soon, at which point I doubt most people will search MySource pages due to the clutter of ill-specified GEDCOM sources.

--Dallan 11:14, 4 April 2008 (EDT)

Works for me. Of course, the concensus opinion of the preferred format (Author Title) for source page titles doesn't work for me at all, but that's because my purposes are, I suspect, different from the majority of users. Any of Dallan's workarounds would solve that, and I'm happy to go the workaround route. For my purposes using an alias to display what I want is a fairly cumbersome approach, so I'll use one of the more direct methods. That's probably going to be a "MySource" item. As Dallan notes, that will probably take it out of the "discovery mode" for many users, but my primary purpose is to present the information content as cleanly and simply as possible. I want something that's fairly clean in presentation, and also simply to use. MySource would be the logical solution. One possible solution for the "discovery by others" problem, is to carefully categorize each item. That's automatic with the standard input fields, but can be added manually as well. Q 11:57, 4 April 2008 (EDT)

Hi Dallan, I understand your reasoning, but I am also in somewhat of a quandary. I envisioned using WeRelate as my "genie" program and downloading gedcoms from WeRelate to another "genie program" to produce reports, charts, etc. However, I had not considered the possibility that the sources generated by WeRelate would create a citation, that I will probably find necessary to edit before publishing a report or chart. Does anyone now have a source titled "United States, Georgia, Walton. Vital Records"? I did not have a problem with the locality order; I created the source for the cemetery and titled it United States, Texas, Comanche, Indian Creek Cemetery.

Would it be possible to allow a user to cite the source in WeRelate but override the created citation and write their own citation?--Beth 20:05, 4 April 2008 (EDT)

Currently we don't generate citations except for the very simple citation displayed on the Person and Family pages. But that citation text wouldn't get included in exported GEDCOM's; instead, we'll include the individual citation fields that are part of standard GEDCOM. It would certainly be possible for you to write a custom citation. The question is where to store it so that it would get included in an exported GEDCOM. One place would be in the "Text" field of the source citation on a Person/Family page. Alternatively, we could add a "Citation" field to Source & MySource pages that could be attached to the source record in the exported GEDCOM.--Dallan 15:34, 5 April 2008 (EDT)

Initial Release of new Source & Repository functionality [20 July 2008]

I finished the new source and repository functionality based upon what we've talked about and just uploaded it. I've tested it and it appears to work, but it's new code so there may be bugs. Feel free to give it a try, and please let me know if I've misinterpreted or forgotten anything. Thanks again for the great discussion!--Dallan 15:34, 5 April 2008 (EDT)

Clicked the "?" and explanatory text came up properly.

Clicked the link for source citations and got no response. Status bar showed that it was trying to load something called "addSource()".Same result with both FireFox and Safari.

Q 17:12, 5 April 2008 (EDT)

Added a new source in page creation mode (typed in the address of the page to be created on a working page, then saved, and clicked the link to the page to be created. That worked fine. Q 11:07, 7 April 2008 (EDT)

Ditto here, pretty much. If I'm editing a page where there is already at least one source, then the "Add source citation" link works fine. But if there are no sources so far, then the Add source citation link does not work, nor does the + under citation ID for events. However, if I click "Add source citation" say, 3 times, then click the +, it puts an S4 in the box -- just doesn't come up with a place for me to add the citation. I'm on OS X (Leopard) and Safari 3.1.--Amelia 18:10, 6 April 2008 (EDT)

Oops - sorry about that. It's fixed now.--Dallan 12:03, 7 April 2008 (EDT)

Can we have the ability to redirect Source pages to Repository pages? Not a big deal, but it could help with some of the clean-up (e.g. Source:New England Historic Genealogical Society).--Amelia 21:22, 10 April 2008 (EDT)

I know we agreed not to have a website "type", but the disappearance of the URL field seems a bad idea. Not only is it annoying not to have a place for a URL for a source that is in fact a website, but in the current setup, the contents of the URL field simply disappear as soon as one hits edit -- unless you know to click on the URL, copy it, then go back to the source page, before editing, that information is just being lost.--Amelia 21:02, 11 April 2008 (EDT)

This is an oversight on my part. If the URL, call number, or repository name field is filled in for a source, I should be adding this information on a repository line. Also, I'll make sure that even if there are no repositories listed, that a blank repository line is shown on the edit page so that you'll have a place to put the URL. I'll also allow redirecting Source pages to Repository pages. You can't simply rename a Source page to a Repository page because the data fields on a Repository page are different than those on a Source page, but you'll be able to edit a Source page and redirect it to a Repository page. These changes should be in place by the end of tomorrow. How does that sound?--Dallan 17:29, 15 April 2008 (EDT)

I thought this was working at one point, but now when I try it, I get "Loop in Redirects" if the source and repository names are the same. If the repository name is different than the source, then it works.--Amelia 15:02, 20 July 2008 (EDT)

Thank-you for pointing that out. I'll fix it later this evening.--Dallan 20:26, 28 July 2008 (EDT)

EVIDENCE EXPLAINED: Citing History Sources from Artifacts to Cyberspace, by Elizabeth Shown Mills [8 April 2008]

From the APG list:

EVIDENCE EXPLAINED Named "Best Reference of 2007"

The April 15, 2008, issue of the national trade magazine, LIBRARY JOURNAL, has selected EVIDENCE EXPLAINED: Citing History Sources from Artifacts to Cyberspace, by Elizabeth Shown Mills, as one of its "Best Reference of 2007" selections. To quote LIBRARY JOURNAL:

"Tracing your family history has become a passion for many people. Mills, an acknowledged expert on archival research, guides you through thousands of different kinds of sources from archives, cemetery records, local and state records, to national government records and web sites. She discusses the fundamentals of analyzing evidence and provides authoritative guidance in citing sources. Every library needs a copy."

Mrs. Mills' 2007 publication, the distillation of years of her own and her colleagues' research experience, has now earned four superlative accolades from the major library media.

Besides its April 15, 2008, mention, LIBRARY JOURNAL initially reviewed Mrs. Mills' work in its November 15, 2007, issue. That lengthy review emphasized that Mrs. Mills' new book, unlike classic citation guides, "focuses instead on nontraditional sources such as militia rolls, frakturs, censuses, and railroad pension files. Recognizing that an increasing amount of those materials is produced and offered online, Mills also addresses e-newsletters, podcasts, databases, and digital images." LIBRARY JOURNAL'S reviewer concluded that EVIDENCE EXPLAINED "is an essential resource for family historians; highly recommended for all libraries."

In December, 2007, BOOKLIST, the official publication of the American Library Association, published its review of EVIDENCE EXPLAINED. The review stated that "the first two chapters are excellent beginning points for the historical researcher, providing in-depth information on the fundamentals of evidence analysis and on the fundamentals of citation." Pointing out that the book contains "more than 1,000 citation models covering print, microfilm, microfiche, Web sites, digital books and journals, DVDs, CDs, audio files, podcasts, and e-zines," BOOKLIST concluded its review with the recommendation that EVIDENCE EXPLAINED "would be an excellent purchase for academic and larger public libraries."

CHOICE MAGAZINE, the official publication of the Association of College and Research Libraries, posted the most recent full review in its March 2008 issue. Zeroing in on the book's value for college libraries and undergraduate students, CHOICE described EVIDENCE EXPLAINED as "a dense guide to citing and analyzing historical records of all types, especially sources not covered in popular guides such as 'The Chicago Manual of Style'." Stating that "the 14-chapter guide could scarcely be more comprehensive or thorough," CHOICE ends its review with this glowing summary: "Overall, this is a key resource guide for scholars and serious researchers who must rely upon and understand historical evidence. Highly recommended. Upper-level undergraduates and above."

We at "Genealogy Pointers" wish to express our congratulations to Mrs. Mills on having her new book named a "Best Reference" book of 2007. Anyone seeking more information on EVIDENCE EXPLAINED, or its author, can do so at the following: http://www.genealogical.com/index.php?main_page=product_info&item_number=3843&NLC-GenPointers1

Reprinted from Genealogical Publishing Company's newsletter, Genealogy Pointers, 4/08/08.

Q 15:21, 8 April 2008 (EDT)

Also, for those REALLY into genealogy...

http://www.ngsgenealogy.org/articles/NGSQVol91Pg260-77-Genealogy&History.pdf--Q 15:34, 8 April 2008 (EDT)

The Cleanup Process [15 April 2008]

We can't redirect pages that are not in the same namespace, so how we handling the moving of "sources that aren't really sources" into other areas? For instance, this page is referring to a genealogy program so it should probably be made into an Article, but since we can't redirect, should we copy over the info to the new page and then delete the old one? I didn't want to delete pages if that's considered "unwiki-like." --Ronni 01:08, 11 April 2008 (EDT)

Please don't feel like you have to spend a lot of time cleaning up sources - at least not the FHLC sources - right now, because I'm planning to write a program to automatically clean them up as best as I can over the Summer. Then we'll probably want to go through and clean up what the program wasn't able to clean up on its own.

I'll make it possible to Source pages to Repository pages or articles by the end of tomorrow.--Dallan 17:29, 15 April 2008 (EDT)

Creating source pages with new format [15 April 2008]

Hi everyone,

Before I create more census source pages by county; it is my understanding that I can do so; even though the automatically created census sources will be designated by the year and state/country region.

Don't we need a "choose" selection for the repository to prevent duplication?

Please check this newly created source page and let me know if I have filled in the blanks as intended. Source:United States. Georgia. Catoosa. 1860 U.S. census, population schedule.

Also should I omit population schedule from the source title? Would you prefer to use this source for other schedules as well such as agricultural and mortality?

I am not sure of the purpose of the title field? What is its purpose? --Beth 11:01, 13 April 2008 (EDT)

Good point on the repository drop-down. I'll add a drop-down list for the repository field tomorrow.

Regarding Source:United States. Georgia. Catoosa. 1860 U.S. census, population schedule, Here are my thoughts:

• You should put commas between the place levels, and a period only between the place and the name of the record set. So I'd rename this to Source:United States, Georgia, Catoosa. 1860 U.S. census, population schedule.

Dallan, I modeled my source list entry after the examples in Mills' new book. The locations are separated by periods when listed larger to smaller and separated by commas when listed smaller to larger. --Beth 21:43, 15 April 2008 (EDT)

• The title field should contain the portion of the Source page title that is the actual title of the source. So I'd probably make this just "1860 U.S. census, population schedule".
• For the places covered field, I'd enter the entire county: Catoosa, Georgia, United States.
• When you add a source by choosing "Source" from the "Add" menu, all three of the above things should happen automatically. I'll bring up this Add Source page when you add sources on Person/Family pages once search is done.
• I'd probably leave the surnames covered field blank for census sources, since there are so many :-).
• I'd leave population schedule as part of the title. I think the other schedules should be their own sources.

--Dallan 17:29, 15 April 2008 (EDT)

Proper punctuation for places in sources titles [31 May 2008]

Quoting Dallan: You should put commas between the place levels, and a period only between the place and the name of the record set. So I'd rename this to Source:United States, Georgia, Catoosa. 1860 U.S. census, population schedule.

Dallan, I modeled my source list entry after the examples in Mills' new book. The locations are separated by periods when listed larger to smaller and separated by commas when listed smaller to larger. --Beth 21:43, 15 April 2008 (EDT)

Comments please.--Beth 08:09, 20 April 2008 (EDT)

Sometimes people who are the recognized experts (Elizabeth Shown Mills), feel a need to make things very elaborate and precise. If one were publishing in a formal journal, you'd probably want to follow her conventions---not necessarily because they are really good conventions, but because the CPG's would expect you to. But in truth, this business of using a "," to indicate small to large order, and a "." to indicate "large to small", seems overly elaborate. I can't imagine anyone other than a CPG would recognize that there was significance to the subtlety. Strikes me as an over-elaboration. Q 10:46, 20 April 2008 (EDT)

As a matter of usability on WeRelate, it's useful to separate levels with one punctuation mark and then to use another mark to signify the end of the place because they alphabetize differently. So if I start typing "United States." I'll get only the U.S-wide sources like the SSDI in the source drop-down, and if I type "United States," I get sources that have been further narrowed. If we always use periods, the sources will all be mixed together in less of a useful order.--Amelia 12:03, 20 April 2008 (EDT)

The point was in reference to using "," to indicate ascending order, and "." to indicate descending order. Or at least that's the way I interpreted Beth's question. As in

"United States.Tennessee.Greene County.Greenville" vice

"Greenville, Greene County, Tennessee, United States.

There's conceivably some utility to that kind of convention. its not intuitively obvious, and I suspect the chances of getting people to a) recognize it, and b) use it consistently approach zero. Q 12:10, 20 April 2008 (EDT)

Thanks to everyone for the comments. I suspect that this method is not just from Mills but the generally accepted manner in properly entering such data. However: if there is good justification for not following the standard because of the requirements of our Wiki then okay. I suppose someone needs to write the Wiki way of properly or improperly <g> documenting our sources. --Beth 19:57, 20 April 2008 (EDT)

That WAS just personal opinion, which I believe is what you were asking for. I believe people sometimes get overly elaborate on mechanics, and focus on "I want to do this right", as opposed to "I want to convey this information effectively". Usually, KISS is the best methodology. and if you KISS something, and your audience understands your meaning, it doesn't really matter how many of "Da Rules" are broken. Of course, if you're writing professionally, then your audience will definitely know the rules, and will not appreciate keeping it simple. Q 20:07, 20 April 2008 (EDT)

Hi Q,

Of course I appreciate your personal opinion. I understand the KISS concept. I just don't believe our Wiki should violate the standards unless there is a valid reason. If using periods inhibits the search function or otherwise compromises the Wiki then I acquiesce.

The concensus appears to be to use commas for various reasons; so in the spirit of the Wiki; I will do so. --Beth 20:45, 20 April 2008 (EDT)

Sorry - I've been working hard on search lately and haven't visited my watched pages in few days.

I was about to agree with Beth that we should change the standard to separate place levels by periods instead of commas because that I think in general we should follow the standard, but Amelia makes a good point: keeping all of the United States sources sorted together is a nice benefit, and I think it outweighs the benefit of following the standard in this particular case.--Dallan 13:01, 21 April 2008 (EDT)

Dallan, when I change the source title from the periods to commas and the page is redirected; my source citation titles are not updated. Why not? I have to manually change the title; the changed title is an indication of which ones I have edited.--Beth 22:28, 8 May 2008 (EDT)

When a source page is renamed, the original title becomes a redirect to the new title. I don't automatically update all of the pages that link to the original title because (a) eventually there could be thousands of pages linking to the same source page and that would be a lot of pages to update, (b) if you follow the original links you're taken to the new title, so it didn't seem that important. You can see which pages link to the original title by clicking on "What links here" in the "More" menu. You'd have to change the source title on those pages manually though :-(.--Dallan 01:44, 1 June 2008 (EDT)

places covered field [21 April 2008]

Dallan, I just created a source page and when filling in the places covered the normal drop down list did not appear. Is this OK? or a bug to be fixed? --Msscarlet1957 12:11, 21 April 2008 (EDT)

It should appear. I just checked and it appears for me. One thing though: it appears only if you're typing into the last line of the edit box. If there is text on the line below where you're typing, it won't appear. There's a technical reason for this limitation. Eventually I should probably turn the multi-line place entry box into a series of single-line entry boxes (like Repository) to avoid this limitation, but I haven't gotten there yet. If you were typing on the last line and it still didn't work, please let me know what browser you are using - that's a bug.--Dallan 13:01, 21 April 2008 (EDT)

Repository namespace [25 May 2008]

Has the repository namespace actually been implemented? Several weeks ago I added Ancestry.com as a repository for some source and Ancestry.com now appears as an article not as a namespace. I also cannot find one namespace using the search feature and keyword Repository:.

I also attempted to add a repository today to a source and highlighting the words add repository did not redirect me to a new repository page. --Beth 09:53, 27 April 2008 (EDT)

The repository namespace was implemented a few weeks ago. Your Repository:Ancestry.com page is one of the repository pages. To add a new repository page, you can either select "Repository" from the "Add" menu, or you can edit a Source page and enter the title of the repository page in the "Repository title" field, then save the Source page and click on the red repository link. Repositories don't have a lot of metadata -- just a place, URL, phone number, and address.--Dallan 13:01, 3 May 2008 (EDT)

I just added a new source Source:Huddle, Rebecca. The 1820 Census for Fairfield County, Ohio with Indexes and the repository fields acted strangely. As I was typing, it kept flashing, almost like it was trying to bring up a drop-down menu, but it never appeared. I'm using Firefox 2.0.

On a somewhate related note, on the entry screen, would it be better to say "Repository Name" rather than "Repository Title"? I think most people think of libraries, etc. as having names and books as having titles. It might throw some people off. --Ajcrow 09:16, 25 May 2008 (EDT)

You're right - I'm getting the same problem. I'll look into it on Monday. Also, I'll change "Repository Title" to "Repository Name".--Dallan 01:44, 1 June 2008 (EDT)

Adding author's name to source title - how to if name is Hemperley, Marion R. [3 May 2008]

Does this work? Hemperley, Marion R. Handbook of Georgia Counties

Note that the period is used to separate the author from the title but in this case this period is used after the initial R. Using 2 periods seems out of the question; so will the one period work properly? --Beth 10:01, 27 April 2008 (EDT)

I agree. One period makes the most sense.--Dallan 13:01, 3 May 2008 (EDT)

Internet sources [1 June 2008]

I know we've discussed this before, but I'm having a hard time wrapping my mind around it. I'm still having a very hard time thinking about not making a source for a website if that's where the data lives. For example, if I find George Debolt in the database Jay County, Indiana Cemeteries, why wouldn't I make a source for that? That is where the data lives.

I have no problem having the source type be the "original" source for websites that have digitized images of records, such as the myriad sites where you can find digitized census pages. However, electronic databases and online record transcriptions are not the same as the original source. The only difference between an electronic database/online record transcription and a book is the medium. If we have a record type "Book," why don't we have a record type "Website"?

If one goes to the page Add a new Source page and wants to cite something like the Jay County, Indiana Cemeteries database, the only way for them to do it is to categorize it as "Miscellaneous." Considering the countless number of "websites with original content" (as stated on "Add a new Source page"), doesn't that defeat the purpose of "Miscellaneous" (which really should be a "catch all" for sources that truly do not fit into any other type.)

I fully realize that people are going to cite whatever in the world they want to cite. However, I think that excluding "Website" as a source type is, in the long run, going to bring more confusion into the mix. --Ajcrow 12:13, 25 May 2008 (EDT)

I do not cite a source if I have not seen it; so I have used Misc. for transcriptions on the US Genweb pages. This seemed to be the consensus of the majority of the active users when we previously discussed this issue. I believe I have a web page source for a cemetery association, that I have not yet changed but I guess that will go under Misc as well. There is no other place to put tombstone transcriptions is there?

Perhaps one could liken the sources on WeRelate to the old subject library index cards. Maybe we could name the Source, tombstones or some such and then cite the web page in the citation? Would that work? I don't know the best way to handle this. --Beth 13:17, 25 May 2008 (EDT)

Further thoughts on this subject. One of the reasons that we decided to eliminate web pages was that they are ephemeral, here today and gone tomorrow. But perhaps we could have a general source for cemeteries; Cemetery Web Pages or something similar. One could then cite this source add the web page in the repository title section. That way we could place all of the cemetery sites under the one source, such as Findagrave; Internment and the independent cemetery web pages. --Beth 15:33, 25 May 2008 (EDT)

I'm a bit of a heretic on this issue (which I think it comes from having both a slightly different view of how sources should be used here, and a certain concern for usability and scalability), but my two cents:

1. I don't think the Jay County Cemeteries database is a website so much as an independently created (non-government) collection of records. It could be available as a paper manuscript, a CD-ROM or a search through a 3rd party, and it would still be the same source page for our purposes.

Precisely my point. We have a "source type" for Books, so why not have one for Websites? Come to think of it, why is "Book" a source type? It seems as if everything else in the Source Type dropdown is a type of source, rather than a medium. What does "Source Type" really do for us? Is it a truly necessary distinction? (Just thinking out loud here) --Ajcrow 18:18, 25 May 2008 (EDT)

1. Similarly, for cemeteries in general, I think each one should have its own source page. If two people have walked it and posted separate transcriptions, note that on the source page. That source page will be much easier to find, and more useful to users, if it is not split between the various transcriptions and uses one common geographical naming convention.
2. Instead of creating a "website" type, which is a whole thicket we've hashed over previously, we could deal with at least part of this specific issue by expanding "government/church records" to deal with other geographically based record sets like cemeteries or independently created databases. In other words, those sources would get predictable geographic-based names.

(Side note/explanation: My general rule of thumb is that if two "things" (for lack of a loaded term) are supposed to be the same -- like a transcription of the census & the microfilm, or the cemetery transcription and the stones in the cemetery -- they should use the same source page. If 99% of the time, the data in the original set will be reflected in transcriptions, I'd rather not design the system based around the 1%. (That's not to say that one couldn't or shouldn't add a notation to a citation on a person page that expresses the equivalent of "I saw this here and (thought) it said X." There are plenty of differences in interpreting handwriting or partial data to be hashed out on an individual basis without regard to what the source page itself is about.) If two things are not the same -- reordered, annotated, revised, etc. -- they get different pages. I never did agree with the "ephemeral" justification, but to the degree it's true, that rule does avoid creating some source pages more likely to disappear.) <p>

--Amelia 15:52, 25 May 2008 (EDT)

This site [4] shows over 48000 cemetery sites online just in the US. There are most likely more online than shown. Just wondering if we should encourage a source for each cemetery site? --Beth 20:33, 25 May 2008 (EDT)

Read Topics 24 and 25 on this page. I had a similar discussion with Dallan about cemeteries and the USGenWeb. I am attempting to find a workaround that meets Dallan's guidelines on this subject but also satisfies my standards for creating sources. --Beth 09:16, 26 May 2008 (EDT)

Jay,

Whether you make a website a source or not depends a bit on what you want "sources" to do for you. There are two underlying reasons why "we source".

A. The first is so that we ourselves can get back to where we found something---perhaps new information has come in and you want to see if what you found has something more to say to you in light of your new information. Perhaps you just want to be sure that your source says what you thought it did.

B. The second reason is that you need to be able to tell others where your information came from. They need to be able to see for themselves that the information you extracted is in fact what's in your source.

The reason why you have to be careful about citing a website as a source is because it may not be there next time you go looking for it. See Link Rot. If your concern is only A, then it doesn't really matter what you cite. If you site "Bob's Family History" site for a key piece of information, it doesn't really matter that a) Bob might remove that information tomorrow, or b) the site may have been taken down tomorrow. In either case you can't get back to the original, but then that's no-one's problem but your own. If that's not a problem for you, siting a website does not harm.

However, if you want others to be able to confirm what you've found, citing "Bob's Family History", is going to be a long term problem for you. So in such an instance, what you probably want to do is look for the underlying sources on which "Bob's Family History" was based, recover those sources, and cite them. If you are really lucky, Bob will have identified his sources for you, and you have an easy road to recovering them. But you probably won't be that lucky because most people don't provide their sources. Some do, most don't. But at least you have the data itself to go on, and often its possible to figure out where it came from. For example, if the author has provided a bit of text, its common for that text to have come from a published work. Doing a search on Google Books will often pinpoint where the data came from originally---or at least a published source.

That said, some web sources can be used with more confidence. For example, if you find something on Google Books, you can expect to be able to return to that work more or less indefintely. Link Rot is not going to be a serious problem with a source like that. Ditto Ancestry, (with the caveat that those family tree's they publish are subject to change without notice, and so don't have good "returnability"---but the census records and other sources on that site are going to be "returnable".) So in such cases, a website as a source is great. Cite it with confidence, and you need a source card in the source namespace to do that on this site. (Actually, if you want, just create a source card in the "article namespace". It will do about as good. Just don't include that "source:" tag at the front.)

But I'd be really leery of citing "Bob's Family History" site. Q 10:06, 26 May 2008 (EDT)

Q, I understand your comments; but we need a final concensus on citing a website within the framework of WeRelate. For example, Indian Creek Cemetery Association has a website online with transcriptions of the tombstone engravings and some actual photos of the tombstones. I have used information from this site for documentation of the birth, death and burial place of several of my research targets. I will probably not in this lifetime visit this cemetery in Comanche County, Texas. Now it is probable that one of my fellow researchers will get there and take tombstone pictures for upload; but in the mean time I intend to cite this web page. Now presently my source on WeRelate is by locality and website. But if you read the other topics I referenced above Dallan suggested that we have a source entitled Cemetery Data or something like that and list the website as a repository. So what I need is another final final consensus; because there seems to be no consensus at the present time. --Beth 10:58, 26 May 2008 (EDT)

Beth, please consider my comments only for their heuristic value. For my purposes, it doesn't much matter how others choose to document their work.

Providing "A SOURCE", is always better than giving "NO SOURCE". The later characterizes about 90% of the family trees' in Ancestry, and probably across the web as well. Of the remaining 10%, most cite a source that almost certainly will suffer from link rot (e.g., Bobs Family History). Perhaps less than 1% of all family trees on Ancestry cite an effective, persistently recoverable source. If its not a problem for someone that a source can't be verified, then its not an issue. I can't say I think it good to recommend bad practice (and citing NO source is certainly bad practice.), but if "rules" are being emplaced, its probably best to base those rules on "best practices", not "best we can do practices".

Q-- That's my point. Best practices state that you cite what you see. If you see a website, that's your source, that's what you cite. Are all websites accurate? No. Are all websites going to be around tomorrow? No. Is it what you saw and what you used to draw your conclusions? Yes. It is misleading to future researchers who are looking at your conclusions if you cite something other that what you actually saw and used. If your citation says something like "Fairfield County, Ohio wills", one would think that you looked at the wills in the courthouse (or perhaps on microfilm). But if what you actually saw and used was an abstract of the wills that were posted on a website (or in a book), stating that you used "Fairfield County, Ohio wills" is misleading. --Ajcrow 16:15, 26 May 2008 (EDT)

As you said, people are going to do whatever they wish to do. I'm guessing that in your work what you are doing is using the site the way it was designed---that is, you are inserting your sources in the appropriate input box. You want to cite something as a source, but its a pretty low value source, and Dallan seems to want to exclude such sources from the Source namespace. Thus you are left with no way to cite where your information.

The site does have a solution for this for this. Instead of treating a personal website as a source, designate the item as "MySource", and let it go at that. The MySource namespace is a fairly elegant solution to this problem. And that is exactly what I think was in mind when that namespace was created---to be used for sources that have limited credibility---peoples GEDCOM's, personal websites, the ephemerata of genealogy if you will. If one allowed such items to be included into the "Source" namespace, you'd basically flood the site with sources of very limited value. This was done initially in the context of GedCom uploads, where about 90% of the sources are ephemeral, but it works equally well in other contexts.

I think we're talking about two slightly different animals here. I agree that MySource could be a good place to put the more-fleeting websites. However, going back to my original example of the Jay County, Indiana Cemeteries database -- that's on GenealogyCenter.info, which is part of the Genealogy Center at the Allen County Public Library in Fort Wayne (definitely *not* a personal website and definitely *not* a "low value source" <g>) It is also a source that is likely to be used by numerous researchers, not just me, which would seem to "disqualify" it (for lack of a better term) as a MySource. --Ajcrow 17:42, 26 May 2008 (EDT)

I sort of wondered if that's what you were thinking, and perhaps should have looked at your example more closely. In anycase, perhaps what you want to cite is the database that they are drawing on/presenting, rather than the website itself.---ie, Jay County, Indiana Cemeteries Database. Q 17:50, 26 May 2008 (EDT)

But how do you separate the two? The database, for all intents and purposes, *is* this website. (Speaking specifically of the pages "Jay County, Indiana Cemeteries," not the entirety of GenealogyCenter.info.) This is the only place the data lives (unless you count a CD of a Microsoft Word file laying around someplace -- and I don't think we'd want to cite that! <g>) --Ajcrow 17:57, 26 May 2008 (EDT)

There are a good many databases on the Allen County Library site, all accessed through the same website, but that's not the only way one could get to the database. It would be possible, if you had the appropriate permissions, to access the computer which houses the database, without ever going through the internet. When you access this site, and drill down to the point where you make a request for some particular bit of information, that information is presented to you at some URL. But the URL is not where the information "lives", its only where its served up to you, and that only for the very specific request you make. The actual data is stored in a database house somewhere (not necessarily locallly) that the server accesses and serves up. As soon as you go away, whatever was presented at that URL is converted to free electrons; the information itself, continues to reside in the database the server drew upon. The URL is very definitely NOT where the data lives.

The correct approach here is, I think, to cite the database as the source. I deal with this problem a fair bit. For example, I draw on Google Books for much of my information needs. When I cite something from a book available trhough Google Books, I don't identify Google as the Source. Rather I identify the Book as the source, provide a bibliographic citation, and then give an electronic source (URL) for that book. If someone else wants to see for themselves what was said in that book, they can click the URL and navigate to the proper passage in the book. But that URL is not the source, only the place the book was presented to me. Same way with the Jay County Database. The source is the database, and you just happen to be able to access that database through Allen County Library, at some particular URL.

I'm well aware of the concept of how data is served up to a browser. (In the interest of candor, I should tell you that I'm the one who converted the data for inclusion on that site. The data is *not* held locally; the only way that the public is going to access the data is through that website.) We may be arguing semantics, but I'm not sure. You seem to be saying that the source is the database; I'm saying the source is the website. If this were a book of cemetery transcriptions, would you say that you should cite the book or cite the underlying data? But even if we were to cite the database, how are we supposed to do that with the current source set-up without it appearing that we went to the cemetery or examined actual cemetery records (in this example)? --Ajcrow 19:59, 26 May 2008 (EDT)

Here's the way Ancestry solves the problem.

Ancestry.com. California Birth Index, 1905-1995 [database on-line]. Provo, UT, USA: The Generations Network, Inc., 2005. Original data: State of California. California Birth Index, 1905-1995. Sacramento, CA, USA: State of California Department of Health Services, Center for Health Statistics. About California Birth Index, 1905-1995 This database is an index to over 24.5 million births occurring in California between 1905 and 1995. Information contained in this index includes: child's name (names may be truncated at 8 letters), gender, birth date, birth county, mother's maiden name (names may be truncated at 8 letters).

Note that they are not citing an url for this, but are describing the Index as an on-line database. And this is the way they think the database should be cited. They also point to the orignal data surce that they used in compiling the database. Q 20:10, 26 May 2008 (EDT)

And I thank you for this discussion. I always find these conversations interesting because they usually help clarify interesting issues for me. Q 17:33, 26 May 2008 (EDT)

If a website citation is going to be used, pointing to a site whose persistence is questionable, then a good choice might be to cite not the website itself, but something like "WayBack", where the material is being stored indefinitely. I've never used it myself, but I understand that "WayBack" does give a decent persistent snapshot of the web at given point in time. It probably doesn't capture everything, but it captures a lot. I noticed this morining that the "Internet Archive" had done a crawl at the end of 2007, capturing something over a billion web pages. So there's a good chance that if the site was up at the time of the crawl, you'd have a good chance at recovering the data. Q 11:22, 26 May 2008 (EDT)

(Adding a new comment, as our thread above was getting indented pretty far <g>)

Good idea! Also, indentions or no, this is getting so long, that its getting cumbersome to get back to the end of the thread, so perhaps starting a new thread altogether would be good. however, I think we are getting to the end of it anyway, so perhaps not.

Let's say that we make the citation like Ancestry.com does. How do we do it? We don't have a source type for website, online database, or plain ol' database and we don't have a suggested/recommended way of titling the page to reflect that it is a database/website without it appearing that we examined something we didn't actually examine. (And yes, I still realize that people are going to do it however they want to anyhow <g>) I'm just suggesting that we come up with something so that we might have a hope of some consistency. --Ajcrow 20:19, 26 May 2008 (EDT)

In truth, what Ancestry seems to want to have done with their instances of this, is to cite them as the author (Ancestry.com). That seems to be about par for the course. When other folks make hard copy books by extracting data out of governmental files, they usually describe themselves as the author, so I guess that works for Ancestry as well. In your case, perhaps its Allen County Library.com, or something of that sort. Perhaps "The Genealogy Center of Allen County Public Library, 26 May 2008. Jay County, Indiana Cemeteries. [on line data-base, accessible at http://www.genealogycenter.info/search_jaycocemeteries.php]. Whether that goes in Source namespace, or repository namespace, I don't know. one or the other. Q 20:35, 26 May 2008 (EDT)

Dallan's Thoughts [1 June 2008]

A few late thoughts (I got affected by the notification-email-not-sent bug and didn't notice this discussion until now; I need to find that bug).

First, the "source type" field is there simply so the system can determine which fields to display and how to title the Source page.

I'd be ok with adding a new "Cemetery/Geographically-based records" source type if the concensus is that it would be useful.

Should we have different sources for different transcriptions? That's one of the questions that started this discussion. Either:

1. we have one source for all (or most -- see below) transcriptions, and the particular transcription that you looked at goes in the "source citation" on the Person/Family page (I could rename the "Text / Where found" field to "Text / Transcription location" to make this clearer), or
2. we have different Source pages for each transcription, and each Source page specifies the source it is a transcription of in the "References/Cites" field.

We've currently chosen to go with option 1. My main concern with the latter approach is source duplication. If we can come up with standards for titling source pages that make it easy for people to find existing pages for the sources they want to reference rather than creating duplicates with slightly-different titles, I think we'll save ourselves headaches later on. The problem with using websites as sources in this regard is that on most websites the author and title are missing or difficult to find, so it will be difficult for people to find existing Source pages for them.

Having said that, we do allow Source pages for transcriptions that are published as books, because the published book has its own unique and well-defined title and author, and so the chance of creating a lot of duplicate sources for the book are small, and most people will tend to use the book title and author as the source title and author anyway.

The Jay County cemeteries example is an example of a database with a unique and well-defined title and author. This particular example seems a lot like a book that happens to be published online, with Allen County Public Library as the publisher. The ACPL is likely to continue to maintain their website. So I wouldn't be against creating a "May, John R. Jay County, Indiana Cemeteries" Source page. I also wouldn't be against creating a "United States, Indiana, Jay. Hillcrest Cemetery Data" (or maybe "United States, Indiana, Jay, Hillcrest Cemetery. Tombstone Data") and adding a link to the Jay County cemeteries page in the source citations on the Person and Family pages.

From what I've seen, websites like Jay County are the exception rather than the rule. I'd rather not add a "Website" source type because that will encourage many people to create Source pages for all websites they find information on, and they'll generally title those Source pages differently so we'll end up with duplicate Sources.

You can find websites that clearly should not be sources. And you can find websites that contain original content and clearly should be sources. And you can find websites in the middle. It seems to me that the general rule ought to be that the URL of the particular transcription looked at should be written in the source citation text field. But online databases with well-defined titles and authors published on well-established websites could be treated as their own Source pages I think. What do others think about this?--Dallan 01:44, 1 June 2008 (EDT)

I agree with that. What I have begun to do is to distinguish between the "original" source that contained the information to begin with, and the "intermediate" source where I captured the information. This serves two purposes: 1) it tells me where I actually got the information; 2) it tells me where they said they got THEIR information. Then, if the intermediate source goes bad (link rot), I've still got a pointer that tells me where I might find the original. But in anycase, the convention I follow is to cite the author of the work, giving him full credit for whatever he's done. For Ancestry databases, where they've extracted information from another work, reorganized the information for data access, I think it's appropriate to give them the attribution, not the original author. But if the work online at Ancestry is simply a scan of the document, where they've added nothing creative, then I don't think there's any reason to cite them as the source. They aren't. They are just the library where I found the work.

I make extensive use of sources like Ancestry. In most cases, making reference to Ancestry as the source makes little sense. I don't for example, explain that I found a particular hardcopy of a book in a particular library---unless it was a very rare document. So if I find a particular book in Ancestry's online library, my first priority is to cite the author of the book, not Ancestry. But since transcriptions (or in Ancestry's case, scans) can have errors, its probably also important to explain that what I looked at was the version on Ancestry. Its for that reason that I include an electronic source in most of the source citations I prepare. But I'm not likely to cite Ancestry as the Source per se---just the place where I got the work---or rather one of the places where it can be found "n-line". I will cite them as the Source if what they've done is a significant reorganization and presentation of the data found in a particular work. Not many things on Ancestry meet that criteria. Q 08:43, 1 June 2008 (EDT)

Thanks, Dallan, that makes sense to me. Is there a way, however, of giving some examples on the "Add a New Source" page? Also, something like the Jay County, Indiana database would still be classified as Miscellaneous, correct? --Ajcrow 09:08, 1 June 2008 (EDT)