Friday, February 26, 2010

Ancestry.com Bloggers Day: Search (Part 2)

Last year I intended to do stupendously rich articles about Ancestry.com Bloggers Day presentations. Since I never got around to it, this year you’re getting my stupidously poor notes.

Today I will finish the presentation of Tony Macklin, the Ancestry.com search product manager.

  • The issues for the design of Search are:
    • Relevance and control
    • Browse (which is very important and also instructional)
    • Speed and efficiency
  • Showed us a prototype for a new, new search
    • (Thomas MacEntee nicknamed it “HD search,” to distinguish it from the old, new search.)
    • The Search Home looked something like this:
      Sketch of the HD Search home
      • You’ll have to excuse my scribblings. Tony was speaking so fast. Hey! It made sense at the time.
      • Tony pointed out that location searching will be back.
    • The Category Search looked something like this:
      HD Search Catagory Page Mockup
      • I was able to mock up my scribbles without too much work because the page was very similar to the new database page look. Don’t expect it again.
    • The basic search form looked something like this:
      Sketch of the HD Search form
      • Usability showed that changing the label from “Birth Place” to “Where did your ancestor live?” was much easier for new users.
      • Another usability addition that is one of those, “Duh! Why didn’t we do that before?” was the addition of a birth year calculator. Enter the person’s age in a particular year to calculate the birth year.
      • Starting out, the search form has just a few fields. But users can click to add fields for family members and life events. Once added, the fields will always be present unless you later hide them. This allows users to configure the search form to match their particular needs.
    • Advanced HD search has transparency about what the search is doing. And it allows users to filter search results.
      • Restrict name matches: Options will be exact, initials on given names, Soundex on last name, and name variants. This filter will be available in the next couple of months, according to Ancestry.com.
      • Restrict to locations: Restrict to state, state and adjacent states, county, or county and adjacent counties. This filter is in testing now which required a change beneath the covers today. That is why the order of results can be different today compared to yesterday.
      • Restrict to national collections only: United States, UK, Canada, and so forth. This filter has been available for about two weeks and is pictured below.
        Advanced search now has collection and record type filters
      • Restrict to record types: historic records, stories and publications, family trees, and photographs & maps. This filter was released in January 2010.
      • Restrict to lifespan. This filter has been available since May 2009.
    • Q. How will you deploy HD search?
    • A. Advanced HD search will be implemented incrementally. Eventually, HD Search will replace New Search.

That completes my report on the lectures from Bloggers Day. Next time I’ll cover Andrew Wait’s closing remarks.

1 comment:

  1. AI, thank you for reporting on the HD Search.

    I can see troubles from "old" New Search repeated and maybe made worse, in part because of the way some underlying databases are indexed but also in part due to how the databases are arranged internally -- as well as the search interface's instruction to the search engine.

    For example, you say "Usability showed that changing the label from “Birth Place” to “Where did your ancestor live?” was much easier for new users." Oh, ugh, eccccch.

    New Users, esp. those new to genealogy, have little idea of the life-paths of their ancestors. They are often beginning with the 3rd quarter 20th-century, when a great many families never moved. Of course current databases will not retrieve the huge numbers who moved from the South to jobs in Northern industry during WWII, and the many migrations during the Depression (city directories for the 1940s-1950s will retrieve many of these people but not with helpful information about where folks came from).

    The switch to "where did they live" will make retrieving relevant data for people who made dramatic moves after the Revolutionary War and through the 19th century more difficult. The "old" New Search system already makes this difficult, particularly due to its lack of sensible sortability either alphabetically or by geographic location **of the record**.

    IMHO, tailoring a search interface to user ignorance is unwise.

    The 2nd major element, underlying search problems, is that the "old" New Search fails to instruct the search engine to retrieve results *for* a given locality **but not** databases that merely have that locality name appearing as a keyword, and **not** a database that has the same word as a locality in its title (as well as in numerous parts of the database).

    For example, searching for someone who "lived in" Delaware will bring up innumerable results containing no matching firstname AND surname from a book about the Forks of the Delaware (Old Northampton Co., PA), one about the Lenni-Lenape or 'Delaware' Indians, several books of extracts of early Ohio records that mention Delaware, a book on early Tennessee settlements where the word "Delaware" occurs, a book on Natchez, MS early settlement (ditto), and so forth.

    Ancestry.com has not elected to fix its "State" groupings to consist in databases **about and/or for** each state, and the underlying indexing problems still exist.

    These underlying indexing problems also include misspellings (putting two "L's" in the name of Vermilion Co, IL; Worchester instead of Worcester for the MA and MD counties, and many thousands more such aggravating errata. The ways the indexing has been done for English localities has really major problems as well.

    ReplyDelete