Incorrect 'Event Place' in indexed collection 'Michigan, Births, 1867-1902'

kenneth reder

See LRZY-RWZ, Louise Mary Elizabeth Jeschke. It is the DQS algorithm that flagged this to my attention, but clearly it is an issue outside of the DQS algorithm. These records are state records for Michigan births and the collection contains multiple counties in Michigan. The 'Event Place' should NOT be set to 'Monroe, Michigan, United States', only to 'Michigan, United States'! Can this be rectified?

Luvezia Bothke, "Michigan, Births, 1867-1902"

Áine Ní Donnghaile

@SerraNola has the lead on the issues with place auto standardization.

Adrian Bruce1

Not merely is there a place name issue for Luvezia, her birthdate appears to be set to 1888 on the Index Record, despite originally being indexed as 4 April 1888.

Are there Auto Date Standardisation issues as well? Or have I already seen something similar in the past?

Julia Szent-Györgyi

Dunno what the bot's aims actually were for dates, but yes, they have been much disimproved as well. You can see it most clearly on index-based legacy data, where the profile has a precise (if sometimes weirdly-formatted) date and place, while the corresponding index entry now has just the year and (in the better case) a country.

I can't really say "what were they thinking??", because they made that clear enough: they wanted to enable searching by specific points on the map or calendar instead of by text labels for said points. The problem is, they tried to do this by moving the text search task from people to a bot, and the bot (unsurprisingly to everyone but FS, apparently) Did It Wrong. And continues to Do It Wrong. Persistently and destructively. So I'm left with the "what were they thinking??" question in regards to the continuance of the clearly-fatally-flawed experiment. Why haven't they reverted the database to its undestroyed state? Did they not keep backups???

SerraNola

@kenneth reder Thanks for reporting. You will find the same problem with the collection Michigan Deaths 1867-1897. The problem stems from one film having one county as the citation so all other counties on the film get the same placename. It can't be edited because it's already correct in the index. Engineers have explained to me that it's a much more complicated process to fix than it would seem, but they are working on it.

Concerning mess-ups in the auto-standardization of dates, ditto—obviously there's no fix yet.

YColeman

Incorrect State For WWII Draft Registration - Should be Montana not Pennsylvania

Serra Nola- Could you please address this. Thanks.

SerraNola

@YColeman This person is shown in search results and in index details as Montana and I don't see anything in the change history. Where are you seeing Pennsylvania?

Áine Ní Donnghaile

@SerraNola The header on the screenshot shows Pennsylvania - one of those mystery titles.

SerraNola

@YColeman Duh…..sorry, I don't know how I didn't see that. The DGS (#102247644) was inadvertently catalogued in with the Pennsylvania Draft Registrations. I will send it to be fixed.

YColeman

Thanks Aine for the green highlight. I didn't know how to do that.

SerraNola

@YColeman Just a side note: There are 66 DGS for Montana WWII Registration Cards incorrectly mixed in with Pennsylvania records. We appreciate every error report as it often reveals a much bigger problem.