New Experimental Feature - Compare to LAC Transcription
Page 1 of 1

Author:  Lindsay Patten [ Wed Oct 08, 2014 5:03 pm ]
Post subject:  New Experimental Feature - Compare to LAC Transcription

I've added a new experimental feature that allows you to compare the transcription of an 1852 page on Automated Genealogy to the transcription on the Library and Archives site.

To access the comparison, first sign on and then open an 1852 page in split view, scroll to the bottom of the page and click the Compare to LAC transcription link. The site will do a search on the LAC site and gather the surname, given names, and age for each record on the page. This make take several seconds as it must do five searches on the LAC site to get all fifty entries. If you return to the page a second time it will load much faster as it saves the names retrieved. The page then constructs two lists: entries that appear on AG but not on LAC, and entries that appear on LAC but not AG. Entries that are the same on both sites are thus left out of the listings.

Unfortunately, the LAC transcription does not record line numbers and doesn't return results in the order they appear on the page so a line by line comparison isn't possible. Instead the page sorts both lists. In some cases this will align the two lists but in many cases differences in the transcriptions cause an entry to appear in a different place in the list, putting other items in the lists out of alignment. Nonetheless, it is usually straightforward for a human to see which lines correspond.

As a convenience, each name in the list of AG names has a correct link that you can use if, after examining the census image, you want to submit a correction. Each line in the list of LAC entries has a details link that takes you to the full entry for that record, although this is of limited utility as the only other column that is included is gender, plus all the location data.

The functionality is somewhat crude but may be useful if you want to get a "second opinion" on what was written. After some pseudo-random sampling I found that I agreed with the AG transcription more often than the LAC version, but this varies depending on the particular transcriber. There are a lot of differences, I haven't found any pages with no differences.

Ideally there would be a way to record that the pages had been compared and any needed corrections added, but I haven't decided if that is worth pursuing yet.

Page 1 of 1 All times are UTC
Powered by phpBB® Forum Software © phpBB Group