Place names
Place names are a recurring problem. Places and administrative units changed over time. Thus, the same place would have to be named differently depending on the date of the event. Unfortunately, however, many users choose one of the names from the list rather randomly.
In addition, place names are used in different languages. Some users adapt the spelling to their own language, regardless of the national language of the place.
To solve this problem I propose an automated cleanup:
Each place is assigned the respective official name with administrative breakdown depending on the date of the event.
To solve the problem with the different spellings I think of two possible solutions:
Either the spelling of the place in the respective national language is used, or the place name including administrative division is displayed depending on the set language profile of the respective user.
As an example I attach the different official names of my hometown Büderich near Wesel:
German names
seit 1975: Büderich, Wesel, Kreis Wesel, Nordrhein-Westfalen, Bundesrepublik Deutschland
1949 - 1974: Büderich, Kreis Moers, Nordrhein-Westfalen, Bundesrepublik Deutschland
1946 - 1948: Büderich, Kreis Moers, Nordrhein-Westfalen, Deutschland
1871 - 1945: Büderich, Kreis Moers, Rheinprovinz, Preußen, Deutsches Reich
1867 - 1870: Büderich, Kreis Moers, Rheinprovinz, Preußen, Norddeutscher Bund
1857 - 1866: Büderich, Kreis Moers, Rheinprovinz, Preußen, Deutscher Bund
1850 - 1856: Büderich, Kreis Geldern, Rheinprovinz, Preußen, Deutscher Bund
1823 - 1849: Neu-Büderich, Kreis Geldern, Rheinprovinz, Preußen, Deutscher Bund
1816 - 1822: Neu-Büderich, Kreis Rheinberg, Provinz Jülich-Kleve-Berg, Preußen, Deutscher Bund
1814 - 1815: Büderich, Generalgouvernement Nieder- und Mittelrhein, Preußen
1804 - 1813: Büderich, Kanton Xanten, Arrondissement Cleve, Rurdepartement, Französisches Kaiserreich
1798 - 1803: Büderich, Kanton Xanten, Arrondissement Cleve, Rurdepartement, Erste Französische Republik
1795 - 1797: Büderich, Erste Französische Republik
1701 - 1794: Büderich (Burick), Herzogtum Cleve, Königreich Preußen, Heiliges Römisches Reich
1417 - 1700: Büderich (Burick), Herzogtum Cleve, Heiliges Römisches Reich
1020 - 1416: Büderich (Burick), Grafschaft Cleve, Heiliges Römisches Reich
Names translated in English
since 1975: Büderich, Wesel, Wesel County, North Rhine-Westphalia, Federal Republic of Germany
1949 - 1974: Büderich, Moers County, North Rhine-Westphalia, Federal Republic of Germany
1946 - 1948: Büderich, Moers County, North Rhine-Westphalia, Germany
1871 - 1945: Büderich, Moers County, Rhine Province, Prussia, German Empire
1867 - 1870: Büderich, Moers County, Rhine Province, Prussia, North German Confederation
1857 - 1866: Büderich, Moers County, Rhine Province, Prussia, German Confederation
1850 - 1856: Büderich, Geldern County, Rhine Province, Prussia, German Confederation
1823 - 1849: Neu-Büderich, Geldern County, Rhine Province, Prussia, German Confederation
1816 - 1822: Neu-Büderich, Rheinberg County, Province of Jülich-Kleve-Berg, Prussia, German Confederation
1814 - 1815: Büderich, General Government Lower and Middle Rhine, Prussia
1804 - 1813: Büderich, Canton of Xanten, Arrondissement of Cleve, Roer department, French Empire
1798 - 1803: Buderich, Canton of Xanten, Arrondissement of Cleve, Roer department, First French Republic
1795 - 1797: Buderich, First French Republic
1701 - 1794: Buderich (Burick), Duchy of Cleve, Kingdom of Prussia, Holy Roman Empire
1417 - 1700: Buderich (Burick), Duchy of Cleve, Holy Roman Empire
1020 - 1416: Büderich (Burick), County of Cleve, Holy Roman Empire
Best regards
Marcus Abram
Comments
-
No automated clean up. Every time an automated process has attempted to fix place names it has just created more problems. The Places database is too incomplete and place names are too complex.
However, your suggestion that the complete history of names for a place be included for each place with appropriate time spans is great and it is in process as the Places database is gradually improved. A recent update in Family Tree helped a lot in that the place type and time span for each version of a place shows both when entering or editing a place name. Unfortunately, improving the Places database is a long, slow endeavor. If there are specific places you need entered or improved, you can request that be done here:
https://www.familysearch.org/research/places
Teaching users to use the correct timespan for place names, as these get added to the database, will be an ongoing project. I have also been disappointed to find that some users just pick the first name on a list of possibilities even when it is clearly not correct. For example, I've seen people enter a birth date of 1830 and then use the first place name on this list instead of the correct one which is last on the list even though the dates are sitting right there:
4 -
The very thought of another automated algorithm applied to place names gives me shudders. No, just no.
2 -
Sorry @Marcus Abram - you're speaking to a bunch of guys who have got badly burned by automated place manipulation. We seem to have lost lots of standardised names held against profiles, leaving us with just the display version. How many? Enough to give us the shakes every time someone uses the word "automatic". Fixing the data loss is still ongoing through the "New Volunteer Opportunity / Help fix place-names" option on the front screen of FamilySearch.
Whether the issue is with the procedural logic or the underlying data, I don't think we'll ever know for certain - FamilySearch might but they're not publicising it.
As for dated entries, that would be a lovely idea but right now, for every area with correctly dated entries, there are (many?) areas without correct (or even incorrect) dates. Until that's fixed...
In all honesty, I'd rather FS somehow produced some training materials that helped people understand what can be - and should be - done with dated entries. I may be hopelessly optimistic over that even...
4 -
Joining the chorus of people who get the heebie-jeebies from the mere thought of "automatic" combined with "placenames" on FS. Automated placename standardization has rendered the place fields basically useless in FS's entire database of indexed historical records, and I don't want anything like that anywhere near Family Tree.
Your basic idea -- of having places labeled appropriately for the time of the event in the language of the person viewing it -- is already in place, in broad outline. However, as Gordon says, the places database is not complete, and fixing that is a highly non-trivial endeavor. You can contribute to it, as he says, using the suggestion buttons in the Places tool.
But even when the database is complete, it will never be as granular as your list: there needs to be a balance between precision and simplicity. The rule of thumb I recall from somewhere or other on FS is that any administrative change that lasted less than five years is too much detail. I would add that most changes that affect only a designator -- such as "Bundesrepublik Deutschland" versus just "Deutschland" -- are also generally too much detail. They're mostly only of interest to historians, and can be safely omitted from a database meant for genealogical purposes.
3 -
Since we're all gathering here to discuss places again, just in case any of you didn't see this posted over in the Places group, FamilySearch is working to explain place names and the Places database:
---- FamilySearch Places: Mission & Purpose ----
@Marcus Abram - before you despair of us completely, do keep in mind that you are never limited to what the Places database contains. You can always make the place names you enter in Family Tree exactly what you need them to be then link them to the appropriate standard.
For example, the "correct" spelling in the late 1700's for the Vikanes farm in Stord, Hordaland, Norway is Wigenæs but the Places database can only use one display spelling per historical time period and Stord has been Stord from forgotten history thorough 2019 when it became Stord, Vestland, Norway. But I can still use the historical spelling and just link it to the entry in the database so the place entry looks like this:
while clicking on this to open the Data View popup shows this:
The hard part is convincing people that this is the correct structure and intended usage of Family Tree and to leave it alone.
This means that if you have someone born in Büderich, Kreis Moers, Rheinprovinz, Preußen, Norddeutscher Bund, during its three year existence, even though it will never be in the Places database, you can still enter exactly that place name and link it to one of the standardized versions on either side the three year span if you feel that is the correct way to enter it.
5 -
It's bad down in Saarland and the Pfalz over the years, switching between France and Deutschland.
I detest "Deutsches Reich" as claimed by FS Place Standards - very wrong years. They still stick to it despite chorus of opposition. (Deutsches Reich covered 1871-1918, not very early years all the way to 1945!)
Norddeutscher Bund is equally bad, too
We the Deutsch prefers the most neutral one - native centric name - DEUTSCHLAND.
2 -
I can understand that the proposal for automatic change of place names leads to some defensive reflexes. Of course, such a profound change would first require a parallel quality assurance project to be set up with a copy of the "real data" so that it is not compromised. A great deal of time would have to be invested in adapting the place data. Only when this is sufficiently accurate and extensive tests have been carried out by numerous users could such an automatic change be applied to the real data. I think the time line required would be several years. But I think the effort would be worth it.
0 -
@Marcus Abram There is an automatic process in place. You can see some of its "contributions" here and here among many other reports of automated placename standardization errors.
0 -
The first requirement for such a process would be for the standardized Places database to be complete. The reason that the auto-standardization of places in the historical records went awry was that the database is far from complete, as you note in your original post. This led to a huge number of marriages in Stord Church, Stord, Hordaland, Norway now showing in the indexed historical records as taking place in Valestrand Cemetery, Stord, Hordaland, Norway for reasons I won't go into here. In Family Tree, the size of the database would make even a 0.1% error rate unacceptable. Even the most cautious pretesting is not going to find all the ways any automatic routine is going to misinterpret places.
So really the best thing will be to get the Places database complete, teach users how to the Places database, and let individual users take care of their own families.
Looking specifically at Büderich which is found here: https://www.familysearch.org/research/places/?focusedId=8918592 you can see that only two time periods out of the twelve time periods on your list that existed for more than five years are currently in the database. You can certainly request that the others be added by writing to: placefeedback@familysearch.org .
Do keep in mind that they have only a small staff. I've had some contact with them and there is only one person working on all of Finland to fix place names. At one point there were four working on Norway but I'm not sure if more than one person still is.
To get off topic for just a second, something I just learned a few days ago is that you can use to Place database ID number to enter places in Family Tree if you are having trouble getting the drop down menu to behave or are not sure if the drop down menu is really showing the correct place. Just find the place in the database, copy the ID number, and paste it into Family Tree. This can be very helpful if two places have the same or similar names in one county:
I though I would mention this in case other people likewise had not be aware of this.
4 -
Gordon Collett,
Don't bother with sending any email to placefeedback@familysearch.org nowadays.
They don't respond promptly, even for a month.
Before the site set up, it was good. No more and they still have not responded to opposition to the wrong place names
And still have not corrected place names connected to the datasets.
0 -
Emailing is still the best way to get a problem on their list. Even if it takes ten years to reach the top of the list.
In the meantime, use the system correctly as it is designed to be used: Enter the correct place name as the free-form user entered place name and link it to the most appropriate reference value from the list of standardized namea so it has an appropriate latitude and longitude.
1 -
You may be interested in the 6 June comment with links to articles on how to request changes or new places. https://www.familysearch.org/en/help/helpcenter/article/how-do-i-request-a-new-place-in-the-database-of-standardized-places
1
