Wednesday, May 19, 2010

Evidence Management in the Wild

Evidence management diagram

See “Evidence Management” for an overview of this series and for links to other articles.

This week I’ll start to evaluate the strengths and weaknesses of evidence management on Ancestry.com and FamilySearch.org.

I outlined my idea of evidence management in two previous articles (“Why Can’t You Get It Right?” and “Evidence Management Explained”) and got some great feedback from you. I had time to incorporate some of it into last week’s article, but if you haven’t read the feedback yet, you should (here and here).

Sorry, Randy. I’m not telegraphing FamilySearch’s plans; I can’t speak for them. I’m writing this series as a challenge to vendors, including FamilySearch. In publicly critiquing products of my employer, I walk the fine line between being helpful and being fired. You’ll notice I never criticize my employer and I never criticize its products without proposing solutions. But I digress… It’s time to walk that line.

Vendor Support for Source Tracking (the Red Boxes)

The table below shows source tracking features that are needed for good evidence management. I’m filling out parts of this table from memory. Let me know if I’ve made mistakes.

A source is a person, document, page, book, web page, or other artifact that supplies information. In the evidence management diagram, each red box represents the information stored by the genealogy website to track a source. It contains information such as a citation, a transcript, a digital image, and a link.

There are two definitions you’ll need to know for this and subsequent articles:

  • assertion – information about a person, relationship, or event. Birth date and birthplace are two examples of assertions. An assertion can exist as evidence or as a conclusion.
  • field – a square area on the computer screen where the user types in information.
Source Tracking Features Ancestry.com FamilySearch.org Footnote.com
Provides digitized sources online 4 billion genealogy records Record Search, IGI, PRF, AF Indexed images of original documents
Can upload digital images Yes No Yes
Links are “hot,” that is, can be clicked to reach destination Yes Yes Yes
Citation templates for referencing offline sources One Two None?
Handles sources uniformly, whether provided, uploaded, linked, or referenced No. Four different ways: 1. provided, 2. uploaded, 3.linked, 4.  referenced Not supported: 1. provided, 2. uploaded, 3. linked No. Three different ways: 1. provided and uploaded, 2. linked, 3. referenced
Manages sources independently of assertions No No No
Supports variety of citation formats/templates for uploaded images None; put in note field Not supported None; put in description field
Transcription field Yes, but inconsistently labeled Not supported No
Annotate images Yes Not supported Yes
Corrections to provided sources Yes No Annotations?
Annotations are searchable No? Not supported No?
Corrections are searchable Yes Nut supported Ditto?

 

Examples

Below is an example of Ancestry.com’s one and only citation template.

  • Because there is only one citation template, it is not possible to follow industry standards for citations.
  • Because there is only one citation template, it favors published sources even though non-published sources are more important to genealogists.
  • As does PAF, Ancestry.com misuses the terms source and citation.
  • While links to records provided by Ancestry.com are easily available from the assertion on the person page, links to non-Ancestry.com sources are nearly inaccessible. One must open the details about an assertion, then open the source, and only then can you click the link. To Ancestry.com’s credit, it used to be worse. Once unburied, the link used to be dead, requiring a copy and paste into a browser. (Click for a larger view.)

Ancestry.com citation format

Ancestry.com handles uploaded sources in a completely different way. Because there is no citation field, one must place it in the description field.

Ancestry.com citation for uploaded source

The New FamilySearch Tree (NFS) has a simple citation template for living memory sources and another for all other sources. As with Ancestry.com, having only one template forces the vendor to favor published sources. The template looks like this:

New FamilySearch Tree source templateNote that there is no field for a link. Also notice that the transcription field, labeled “Actual Text,” is appallingly small. While Ancestry.com has the annoying practice of ignoring line breaks when displaying transcriptions, FamilySearch takes the annoyance to the extreme. NFS runs together the entire citation template. As a result, the citation above is displayed in this incomprehensible format:

NFS Source Display I’m out of time. I’m back to genealogy the rest of the week. Next week I’ll move on to the heart of evidence management, the green “Evidence Box.”

3 comments:

  1. Just as an aside - re "As does PAF, Ancestry.com misuses the terms source and citation", I would find it useful if you could put up the "correct" definition of those two alongside the PAF and Ancestry definitions, so I can understand the differences. I don't use those 2 for recording stuff but do use a GEDCOM based program, which may, or may not, have similar "issues".
    (I put "issues" in quotes not because I don't believe you, but rather because on my side of the Atlantic this sort of thing doesn't register anywhere in Family History. On the other hand being trained as a mathematician and having done data modelling, I do appreciate robust definitions)

    ReplyDelete
  2. AI, very interesting exploration.

    I think it would be useful for continuity to make plain that the element of the Ancestry.com site that you are talking about is making citations to elements of Member Trees, and on footnote.com to the Individual Pages.

    There is an underlying problem with the way Ancestry provides automatic links to some database items together with some sort of citation.

    In your table:

    Handles sources uniformly: provided sources, uploaded, linked, or referenced
    --No. Four different ways: 1. provided, 2. uploaded, 3.linked, 4. referenced

    The automated linking is not to the image of a document (where such an image exists), but to Ancestry.com's extract of items purportedly from that document (Ancestry.com persistently calls these extracts "the record") -- such as a US Federal Census enumeration for a given individual or household.

    One of the many problems with these extracts is that often the extractor/indexer invented relationships among members of a Census household that are not given in the enumeration itself. For example, in 1880+ enumerations if a person is 'grandson' or 'granddaughter' to head of household, the extractor/indexer picked some other adult in the household to encode as a parent of the grandchild. The enumerations almost never specify such a relationship, and an actual parent of the grandchild often is not in the household.

    To create a citation for the actual image is a manual process. It is also possible to create a link to the image's URL, but there is no way to add a citation for such weblinks -- only a name for the link.

    ReplyDelete
  3. I have the feeling that what you are saying is important, but you lost me in your posts. Can you say it in laymans language?

    ReplyDelete