Wies, Melk instead of Vienna (Austria) - extensive recurrent transcription error in event place
Hi ,
I've been bumping into a recurrent transcription error in which the town of Vienna, Austria is replaced by Wies, Melk, Austria. I looked it up and Wies is a tiny hamlet/village with a population of less than 100 people, ~100 km from Vienna...
This error is quite frequent, I've seen it not only involving entire pages of records but what it looks like entire volumes at times. I'm finally posting about it today after yet another occurrence:
Please see this https://www.familysearch.org/ark:/61903/3:1:33S7-9B2Q-9TX?view=index&action=view as an example, including adjacent pages.
Is it possible to mass edit those wrong records for entire pages/volumes? Maybe even with the help of automation / script to be used by some techies / bots?
Theo
Answers
-
@N Tychonievich seems one for the list of placename errors.
Please and thanks.
@Theo Rafael Is a known problem of the new AI. It has a great imagination🙄 and cannot be resolved easily.
0 -
I myself have two very simillar examples and would love to know, where to report them.
First one includes village of Batorowo (then Battrow) and town of Złotów (then Flatow) in nowadays Poland. In one catalog thre are four (?) groups of photos, first one from Battrow (Evangelical church), the rest from the Catholic church in Flatow. All of them at "place of event" are fixed at Battrow (DGS 008021238)
The other includes an error made while taking photos: village parish of Bahrendorf, Kreis Briesen (nowadays Niedźwiedź) is listed as and attatched to Barendt (nowadys Boręty), 150 km away (DGS 008111346)
0 -
@Theo Rafael, noooo, automation is the cause of the mess!
Although the bot seems to be hiding its activities here: there was no parenthetical "original" on the index detail pages that I looked at (e.g. https://www.familysearch.org/ark:/61903/1:1:C4B8-7MMM).
Add to the mix the fact that "Melk, Wies" is the closest "match" that the Places database has for the "Austria, Niederösterreich, Wien" that's apparently in the Catalog database for Vienna, so currently, the only way to get the Catalog to cough up an actually-categorized list for the city is to search for Wies (https://community.familysearch.org/en/discussion/comment/543045#Comment_543045).
Gaah, what an utter disaster. At this point, I don't believe FamilySearch will recover from it in this decade, given how they continue to deny the vast scope of the problem, and keep poking their fingers into the little holes in the dam that's leaking over the top.
3 -
Thanks, interesting.
In any event this would be a classic case for automated script. In this particular case it should be a no brainer to just involve a strict automated script because as I said Wies is a negligible village of about 60 persons while Vienna obviously has thousands or tens of thousands of records I would imagine...
If instructed correctly, every location which is marked as Wies would be automatically changed to Vienna.
If there are any registrations from Wies itself these could be either blocked from the bot or we'll just have to sacrifice those few as a small price to pay in order to fix the much larger issue of Vienna records which is several orders of magnitude larger at tens of thousands I would think.
I don't have experience with the inner workings of FS and don't know if such a script is even possible but it would be a reasonable/ logical way out of it..
Theo
0 -
@Theo Rafael The problem is an existing automated script that created the issue in the first place.
3 -
There's a FT data quality related session scheduled at RootsTech (although disappointingly it's listed as in-person only & without a recording). I imagine this whole standardised placenames issue will be a major topic of conversation there. @Theo Rafael you may be interested in this relevant post: https://community.familysearch.org/en/discussion/156991/familysearchs-tree-integrity-initiative#latest
0 -
@Theo Rafael Thank you for your report. We have passed it on to the appropriate team group. Unfortunately, they are a small group with a large backlog, so we cannot predict how long it might be before you see a correction.
1
