The following news release was received from Matthew Deighton at Ancestry.com:
Ancestry.com, BehindTheName.com, and WeRelate.org announce an improved approach to finding variant names in genealogy searches. Up to now, most genealogy websites have had to rely upon Soundex to return variant names in response to searches. These approaches often miss variants that should be returned, or include variants that aren’t very similar.
Ancestry.com, BehindTheName.com, and WeRelate.org have created an open-source database of name variants that is free for any website or genealogy software developer to use. Tested against pairs of names provided by Ancestry.com, it reduces the number of missed name variants by over 25% in comparison with Soundex.
How you can help: A large portion of genealogical expertise involves learning variant spellings for the surnames in your tree. Why not share your knowledge with others? By adding your variant spellings to the database, searches on any website that uses it will include your variant spellings automatically. You can review and add variant spellings here: http://www.werelate.org/wiki/Special:Names
In addition, we need people to review the changes that others have made to the database, to make sure that we have multiple pairs of eyes reviewing the names that are being added and removed. You can review changes that others have made here: http://www.werelate.org/wiki/Special:NamesLog
If you are a website or software developer: The database and source code are available at: https://github.com/DallanQ/Names
In addition to the database of name variants, the source code also includes a function to return the similarity score between any two names. This function has been found useful in duplicate detection.
More information about the project can be found at: http://www.werelate.org/wiki/WeRelate:Variant_names_project
While I welcome the announcement and see the end product as a certain improvement over what we have now, my visit to the site strongly suggests to me the process may need improvement.
For example, using my own surname as a “case study”, there were examples of variants someone had apparently added, but with no explanation of why even though space is provided for such to be offered.
I would very much like to know why the specific examples were added as I do not believe they should have been. Without knowing whether the “additions” were indeed added by someone who is not formally part of the project, rather than being a “template” or even a “glitch”, and what reason was offered for the additions if they were “contributed”, I am very reluctant to offer examples that were not included for which I can cite documented proof.
I have already contacted WeRelate and asked for a clarification. Will gladly share the answer I get with everyone, perhaps doing so in my own “column” at Area-Info.