Weird search results

Ros Haywood

For the past couple of weeks, I have noticed an inconsistency between transcriptions and search results. ALL of these examples have correct transcriptions, correct images (obviously), yet the search results are different. Examples are:

Individual lives in Molland, Devon, England - yet their residence is listed as Village, Guadeloupe

Another individual lives in Tormoham, Devon, England - yet their residence is listed as Union Township, Scotland, Missouri, United States

Families living in Millbrook, Cornwall, England - they are listed as living in South Millbrook, County Down, Ireland (while their siblings are correctly in Cornwall)

Family lives in Stoke Damerel, Devon, England (where the 1841 census is taken) - yet their residence in the search results is listed as Villa, Nord-Trondelag, Norway

Family lives in Whitestone, Devon, England - yet their residence is listed as Farm, County Cork, Ireland

What is going on? Why are the Search Results different from the records? Different countries, even?

Ros Haywood

https://www.familysearch.org/search/record/results?count=100&offset=100&q.givenName=Elizabeth&q.residenceDate.from=1841&q.residenceDate.to=1841&q.residencePlace=Devon%2C%20England&q.spouseGivenName=William&q.surname=Beer

This is the link to the Elizabeth Beer whose residence is Village, Guadelupe; and the Elizabeth Beer who was living in Tormohun, but has the residence of Union Township, Scotland, Missouri, United States .

Gordon Collett

Can you post links to these search results? At first glance this seems to be a different problem than the previous reports of incorrectly autostandardized names. So I'm wondering what is going on.

Julia Szent-Györgyi

Is the search display automangling ::ahem, sorry:: autostandardizing "on the fly"?

Ros Haywood

https://community.familysearch.org/en/discussion/comment/436253#Comment_436253

I can't really. They are Search Results pages, so I could link you to one of the pages, but you would have to look through a hundred or so entries to find the weird ones.

Julia Szent-Györgyi

I searched for any place = Molland, Devon, England, and filtered the results for residence in South America: https://www.familysearch.org/search/record/results?count=20&q.anyPlace=Molland%2C%20Devon%2C%20England&c.residencePlace1=on&f.residencePlace0=8

On the search results page, the first match lists a residence of "Lime, Las Carreras, Sud Cinti, Chuquisaca, Bolivia":

However, there is no sign of Bolivia on the index detail page (https://www.familysearch.org/ark:/61903/1:1:JCRM-SPV). It just says "Lime Regis":

In other words, there is automangling being applied and used in the search but not being shown in the index detail view. This is Very Bad.

Gordon Collett

@Ros Haywood , these are actually quite easy to find. And you can post examples by posting search URL's like this: https://www.familysearch.org/search/record/results?q.residencePlace=Union%20Township,%20Scotland,%20Missouri,%20United%20States&f.recordCountry=England

To get here, I did a country limited search of England and put as residence one of your examples of incorrect place and got this:

Scanning through a few of these shows the common factor. The residences are listed as "Union St." So this is another wrinkle, a less obvious one, in what is turning out to be basically a universal problem of auto-standardization not working for incomplete place names that are not listed in the Places database. What I don't understand is how in this situation Union St. got translated into Union Township in Missouri.

You can read about the more typical situation here: https://community.familysearch.org/en/discussion/comment/435802#Comment_435802

Lime Regis is one of these more typical examples because if you type just that into a place data field in Family Tree you get:

I'm not sure why, with the autostandardization, they don't have both Residence and Residence (Original) since they have Event Place and Event Place (Original) in these census records.