ERROR REPORT - Auto standardization
Again - record set New York, New York City Municipal Deaths, 1795-1949
https://www.familysearch.org/ark:/61903/1:1:2WH8-B99
The street, Stanton, has been indexed in the Address field. Search results report the infant died in the Nebraska town of Stanton.
Please escalate to the Engineers for correction.
Thank you.
Answers
-
I know the mods have asked for separate threads for each instance, but I have several more from this same record set in the same time period. https://www.familysearch.org/ark:/61903/1:1:2WJ7-JV5
The burial was in Lutheran Cemetery, now known as All Faiths. The burial place has been corrupted to a location in Germany.
0 -
Residence at 53 Suffolk changed to a location in England.
0 -
A birth place abbreviation for Philadelphia has become Finland.
https://www.familysearch.org/ark:/61903/1:1:F6MN-FQ4
0 -
0
-
And another address on a numbered street becomes China.
0 -
Death in the NY Alms House moves to Massachusetts.
https://www.familysearch.org/ark:/61903/1:1:2WH8-RJX
0 -
0
-
0
-
0
-
And another numbered street becomes São Tomé and Príncipe. https://www.familysearch.org/ark:/61903/1:1:F6MN-X4N
0 -
0
-
-
0
-
0
-
Another numbered street - but now we moved to Cameroon. https://www.familysearch.org/ark:/61903/1:1:F69X-ZM5
0 -
0
-
0
-
0
-
0
-
0
-
0
-
0
-
This one not only moved out of state, it moved back in time to British Colonial America.
https://www.familysearch.org/ark:/61903/1:1:FD1B-LZL
0 -
A shift from Germany to France for the place of birth. https://www.familysearch.org/ark:/61903/1:1:FD1B-1ZP
0 -
I was searching a very narrow time frame - for children born between 1850 and 1870 and also died between 1850 and 1870, in Manhattan. The surname was also very limited, with wildcards. I had fewer than 200 results and more than 20 showed problems with the autostandardization.
1 -
I admire your persistence here - I think the engineers must have grasped the scale of the general problem by now - but other than perhaps getting the examples addressed relating to those we are personally reporting, where is this leading?
I've given up on providing such examples (well, maybe I'll be tempted in future by a particularly annoying one!) as I can't see the overall problem ever being resolved. If you remember, some months ago it was even acknowledged that it was anticipated (by the engineers) that perhaps millions of such errors would be created, but this would still be a relatively small part of the billions of records in the auto-standardization exercise.
For the likes of you and me this is not too great a problem: through experience, we are still locating the sources. The problem is of far more importance to the inexperienced users who, because of the scale of the problem, are completely missing out in identifying the records that are on FamilySearch and would prove of great value to them.
I think the problem will never go away as long as the engineers view this exercise as a "success" (say from a "mathematical" standpoint) and will never acknowledge that has all been a big mistake.
Again, I admire you, Gordon and others who continue to draw these examples to attention, but just don't see spending time on this vast and never-ending problem as a sensible use of my valuable time.
1
