Has a bug led to place names that had already been standardised needing to be standardised again?
Answers
-
Paul said: As there are a number of current issues relating to standardizing I thought I'd add another example I came across just now where some "background" work appears to have caused problems with inputs being no longer recognised as having been standardized.
Any chance of FamilySearch engineers sorting this, or do users have to go back and "re-standardize" all the "undone" examples we find ourselves?
Oddly, the marriage data found here has not been affected by this problem.
0 -
Jeff Wiseman said: Paul, although you can get at it and view it through the person details page, remember that the "marriage data" that you are talking about is actually NOT stored as an attribute of the Person's record. There is a thing called a "Couple Relationship" thing (i.e. a different type of record) where that is all stored, and then that Couple Relationship record is attached to connect the the two person records that it relates to.
If a manipulation tool for place names was run through all the person records in the database, it would totally miss all Couple relationship records in the system, along with their events (including marriage events).
In Information Modeling parlance, this is called an "Associative Entity". It is a discrete Entity (viz, a couple relationship record) that "associates" 2 other entities (viz, person records) with each other.
In other words, a marriage date and place is NOT part of the person records for either a bride or groom. It is part of a "Couples" record that connects the bride and groom together. This is part of the essential model of a genealogical record.
Although some systems may oversimplify this and implement things without the separate associative entity (i.e., the couple relationship doesn't formally exist in the implementation), they do it by having to maintain the same data (e.g., marriage date) in two separate places (i.e., in the bride's record and in the groom's record). So inevitably you wind up with synchronization problems between the two records that need to remain the same.
So when you implement a product, the closer that you structure your implementation to the essential relationships in the system, the better things tend to turn out. FS has done this for the couple relationships. That's why it is totally plausible that a clean up on place names for person records would not affect those for relationship records.0 -
Adrian Bruce said: Now I am really perturbed to find that Paul has still got "No Standardized Place Name" because if I go to "Improve Place Names" and ask to do Names for the United Kingdom (or England) - there are NO place names for the UK or England to be improved.
Perhaps naively I assumed that the volunteer effort had therefore fixed the British Isles. Instead, it looks like the filter mechanism is not finding the unstandardized names in the British Isles. (And yes, I am aware that there may be a chicken and egg situation if the original place names are rubbish but Paul's aren't)
I guess that I need to review my own stuff again....
So can FS please explain why I didn't see the option to work on United Kingdom place names? I've tried it in the browser and in the Android app.0 -
Paul said: Thanks for that detail, Jeff. Still hoping that there might be an admission / explanation on the general issue from a FamilySearch employee on the likely cause of this problem.0
-
Jeff Wiseman said: I guess the takeaway from my comments would be that whatever "fix" they applied in the system not only had problems, but it was INCOMPLETE as it ignored standard place assignments in other non-person record types.0
-
-
Adrian Bruce said: Well, there we have the clearest proof yet that you really did standardise the place-name because - unlike me and probably most of the rest of us - you actually wrote "Standardized Place" in the "Reason".
But the standardised value that you set has been ripped away by whatever background utility FS ran for whatever reason. (A utility that I believe that FS has not admitted to, nor have the reasons been admitted to).
Possible reasons include that it was date inappropriate (1826 but no "United Kingdom") but also that particular one is a mish-mash of alternate names (that's down to reality, not FS - the "correct", standard name in FS is "Horsham St Faith with Newton St Faith, Norfolk, England, United Kingdom")0
This discussion has been closed.