Wednesday, January 29, 2014

Year of the Obituaries

2014 is FamilySearch's Year of the ObituariesFamilySearch has proclaimed 2014 as “The Year of the Obituaries.”

“New U.S. obituary indexing projects have appeared over the past few months and more are coming,” wrote FamilySearch’s Katie Gale. “As FamilySearch.org continues to focus on modern records that connect recent generations, these obituary records are going to be invaluable.”

I think focusing on recent records is a good way to engage new genealogists. Often a new, potential genealogist begins their search by typing in their own name and then their parents’ names. Providing recent records increases the chance that they will find something in these first searches. Obituaries, in particular, may have details beyond vital facts, and these details may have a greater ability to catch their interest.

Obituaries may also document recent vital events for which government records are not yet available. As concerns about privacy increase, government records are becoming less and less accessible.

But I must confess that I’m not sure FamilySearch’s indexing tool is optimal for indexing all the names in an obituary. It looks to me as though they are trying to apply a tool, built for structured information, to a record without formal structure. To accomplish the task, FamilySearch is indexing each name in the obituary as a separate record and indexers are expected to insert additional lines or mark extra lines blank to accommodate the variable number of names.

Don’t get me wrong. I love the fact that they are indexing most names. I just worry that the change in paradigm may be confusing to indexers trying to move between structured records and obituaries.

“indexing obituaries can be a bit tricky,” wrote Gale. “They can be deceptively difficult to index,” states project instructions. In addition to the standard training materials for indexers, FamilySearch has produced additional training materials.

Without a specialized tool, users must learn complex rules. For example, “Mrs. Dan (Terry) Archer” is indexed one way, “Dan (Terry) Archer” is indexed a different way, and “Marvin (Tiny) Beers” is indexed a third way. Consequently, not everyone mentioned in the obituary is indexed, but handling all the nuances is beyond the ability of a generic indexing tool.

When youth groups come to my family history center to learn indexing, without exception their adult leaders object to giving any training before throwing youth into the fray. “We have adults to help them.”

“Have the adults been through the training?”

“I don’t think so.”

Last Wednesday’s group could not find any easy indexing projects, so well-meaning leaders directed several to obituary projects before we could tell them otherwise. I winced.

I’m also disappointed that FamilySearch is not indexing most death dates. “Most obituaries don’t include an exact death date,” wrote Gale. “Don’t try to determine which date is meant by statements like, ‘He died last Wednesday.’ If a death date is not specifically indicated, use the most recent date on the document, which is often stamped on the card.”

That’s a severe limitation that robs the resulting historical records with the most critical piece of information. FamilySearch should have indexers enter exactly what is said about the date (such as “the 14th, inst.”) and afterwards calculate the date.

I did like the fact that the indexing tool provided a different set of fields to index for the deceased than for others named in the document. This would be a central feature in any indexing tool specifically designed for obituaries and other non-structured records. And the lack of an exact death date may not be all that bad. FamilySearch’s search system is incapable of searching exact dates anyways.

Obituaries are a great record and FamilySearch’s English speaking indexing volunteers are optimal for indexing unstructured English records. Neither OCR nor non-English indexers are up to the task. This is a great campaign and I’m glad FamilySearch is taking it on. We should all pitch in and help make 2014 “the Year of the Obituary.”

6 comments:

  1. Hi Insider,
    I'm indexing this project and you are right, it is complicated. On the plus side, the "how-to" materials are the very best I've seen in my 5 years of indexing, and I recently used the Live Chat feature to get a quick answer to my question that wasn't covered in the training materials. Thanks for highlighting the project!
    Sara Scribner

    ReplyDelete
  2. I have indexed over 175,000 names, but the obituaries are complete being done wrong in my estimation. I don't intend to index any more of them, until they straighten out the process.

    ReplyDelete
  3. There are a number open source technologies coming out of Internet search companies like Google that would be much better suited dealing with this unstructured data. It shouldn't be much different than indexing websites for search. It would seem to me that using one of these technologies along with transcribing the obituaries would be a much better approach.

    ReplyDelete
  4. I have just about stopped indexing after years and years....the problem is the obits, utterly worthless in my opinion. I have indexed over 195,000 names, but no more until they give us something decent to index.

    ReplyDelete

Note: Only a member of this blog may post a comment.