Pages

Wednesday, February 3, 2016

#RootsTech 2016 -- Tough challenges which need automated solutions

Session taught by Diane Loosle, Senior Vice President of Patron Services at FamilySearch, Director of the Family History Library and both a certified and accredited genealogist and David Rencher, AG, CG, FIGRS, FUGA, employed as the Chief Genealogical Officer for FamilySearch. A professional genealogist since 1977, he is an Accredited Genealogist CM with ICAPGen SM in Ireland research and a Certified Genealogist SM with the Board for Certification of Genealogists®. He is a past-president of the Federation of Genealogical Societies.

The Innovator Summit is an opportunity to listen to subjects that are not usually discussed in genealogy conferences and certainly not the subject of much discussion among entry-level genealogists. It is also a small window into a world that most of us rarely have contact with who are neither software developers or work for large genealogy companies. The rooms at the Salt Palace are set up for the huge classes planned for the #RootsTech portion of the conferences and it makes it look like the attendance is very light. In reality, it is the segment of the market that is small and the classes are a good representation of the interest.

The purpose of the presentation is to advance the issues that need to be resolved though innovations in family history.

First is the need to have a word space between the raw data and the conclusions. We need a work space that is apart from the FamilySearch.org Family Tree. Some use spreadsheets. I use both spread sheets and Google Docs documents to hold information while it is being evaluated. We need tools that can help in this situation to narrow the possible related individuals in a record such as a census.

There should be lists of name variants used in a search. As we search records, there may be discrepancies. Our analysis consists of comparing several records of different types that may resolve the inconsistent or incomplete data. This comparison process is labor intensive and could partially be assisted by software that would provide a structure for making these types of comparisons.

Example of comparing census records for various years by extracting similar surnames from the same area to determine relatives. This is an example of moving beyond simply looking at the data and analyzing it for patterns. Computers do well with patterns and so this process should be amenable to computer analysis.

5 comments:

  1. Call me a Luddite but why does that need an automated solution?
    If everything in family history was automated there would be no point in doing family history.
    If one could simply log in and access ones family history compiled and certified automatically by computers what would be left to research?

    I understand the point of making records more easily available to people by using technology.
    I understand the reasons behind offering the researcher the chance to view a transcript and a facsimile of the original.
    I understand the reasons behind the compilation of recording “transient” artefacts such as tombstones that could be removed eroded or destroyed over time.

    But I cannot understand why there is a compulsion to analyse and make a deduction of the facts supplied.
    Surely that defeats the whole object of the hobby?

    Cheers
    Guy

    ReplyDelete
    Replies
    1. You raise some rather interesting issues. I am compiling a list of topics for blog posts and I will put your comment on a post for my comments. Sorry I don't have time right now with RootsTech going to respond more fully.

      Delete
    2. I totally agree with Guy - computer or human arbitrator is a dangerous game to arrive at the truth. Take the example of a female ancestor, often given a middle name, yet after 40 years of reading documents, I have never seen so much as a middle initial. What decision would the computer / human make? BTW - that middle name is the same as the surname of a supposed maternal grandfather.

      Move back in time before census, when records are more elusive - just because I have searched and never found proof & I list "wife unknown" - while another claims a certain wife. Will my negative records be checked, my reasoning analyzed and/or will the same be given to the certain wife advocate.

      This requires a huge amount of research time for just one person in a world tree of millions, nay billions.

      What if I disagree with the final judgment?


      Delete
  2. Although the word "determine relatives" was used in the last paragraph, I wonder if "find possible relatives" is more fitting. I think the problems proposed and request for automated solutions are really to help us find possible records to evaluate more easily and help us work through them. I don't think the automation will ever truly be able to make conclusions for us. Otherwise what would a work space be needed for?

    ReplyDelete
    Replies
    1. You are very right. I should have sail determine possible relatives. Sorry we didn't get to visit longer at RootsTech.

      Delete