Name standardization for localities with insufficient presence in FamilySearch database

+14 votes
522 views

We now have the FamilySearch Place Name database established for our location fields. As many locations are only insufficiently represented in this database do we want to discuss standards for naming such localities?

Let's discuss Germany and start with Aachen. It's in WikiTree as the birth and death place of Charlemagne. Forget for a moment that his birth place is completely unknown, his death place is considered certain. On Charlemagne's profile it reads Aachen, Rheinland-Westfalen, Deutschland. The oldest description in the FamilySearch database when German is selected as language is Aachen, Aachen, Rhein, Preußen, Deutschland.

  • Deutschland (Germany) did not exist yet, the town was in the Fränkischen Reich (Regnum Francorum, Royaume des Francs, Kingdom of the Franks, Francia)
  • I suppose the second Aachen is supposed to denote a county. There is no county Aachen. Aachen is "kreisfrei" (does not belong to a county). In any case, counties would not have been in existence in the 8th or 9th century.
  • Rhein is supposed to be the Prussian province, it was created in 1822.
  • Preußen could at the earliest be referred to after 1525.
  • Rheinland-Westfalen is not a political entity, the Rheinland it is a region within Nordrhein-Westfalen, a state that was established in 1946.

So, what do we call Aachen in 814? 

WikiTree profile: Charlemagne Carolingian
in Policy and Style by Helmut Jungschaffer G2G6 Pilot (604k points)

Feels this is a big step in the right direction. Is there any register you can access to see whats in it?

I got a feeling Swedish parishes is not part of the name which is a little bit odd as most genealogy people use parish eg. 
Yngshyttan, Färnebo, Värmland, Sweden 
is suggested to be 
Yngshyttan, Värmland, Sweden

Färnebo is the parish and normally used in the name 

@Magnus: I see a lot of these Swedish place names with parish omitted. It is not just "odd" but quite a problem whan you are sourcing or correcting profiles, because the parish is where the records are.

Yngshyttan is at least a unique name (I think?), but what would you make of stuff like Åby, Östergötland, Sweden or Hulu, Jönköping, Sweden?

I feel its a big step that WikiTree validate and and get a country for all profiles but odd that the FamilySearch Place Name database doesn't have parishes as they are good documenting the Swedish Parishes....

As WikiTree got a new language field maybe more field will come soon. I am getting more and more convinced that you should do the same with location names and connect them to Wikidata as we have done with profiles...

Having a name standard with the parish in it makes life easier. 

I saw that Arkiv Digital latest version now have County letter in the citations which is a step in the right direction plus life adding Wikitree categories will be easier.... If we would have used Source templates then we could have had Categories set automatically..... I play more over at Wikipedia land and they use templates to set maintenance categories ==> you easier find profiles that has an error etc....

New features new problems in the never ending life of genealogy ;-)

The FamilySearch Place Name database must contain place names entered by transcription volunteers, like the diligent producers of Ancestral Files.

Working on correcting the -ssons who should be -dotters I find large batches with this origin. There is a combination of utter devotion to scouring the books for every sibling and unnamed wife back in the 1600s, wilful ignorance of the naming practices (always using -sson even when the name is in plain sight as -dotter) and utter disregard for geography, resulting in these cryptic parish-less locations. With luck there will be others in the same family network having a parish-only location, so the whole thing can eventually be figured out.

While a tool for standardization of locations is good, I just don't trust the FamilySearch Place Name database very far.

The odd thing is that some part f family search has documented parish Färnebo  and Yngshyttan a place in this parish...

I guess the name database is this
https://familysearch.org/stdfinder/StdPlaceLookupResults.jsp 

Then Yngshyttan is
Higher level places: Värmland,Sweden
Name variants yngshyttan
Place type: Type: Populated Place
                     Code: PPL
                     FC: P

Place time period: 1779-
Standard text: Yngshyttan, Värmland, Sweden
Culture: Sweden
Iso code: SE-S
Geo code: 59.75 N 14.2333333 E
Identification number: 3516589


So I guess the standard text is the wrong part that should be 
Yngshyttan, Färnebo, Värmland, Sweden

I hope WikiTree store the Identification Number with the location .... 

 

I wonder if WikiTree stores the FamilySearch Place Name database at all. Or if it is just linked to, wherever it is.

Did you notice the exchange about "Acton, Middlesex, Massachusetts, USA" coming out as out as "Acton, Middlesex, Massachusetts, Amerikas Förenta Stater" in the other thread: https://www.wikitree.com/g2g/330344/england-united-kingdom-other-automatic-placename-suggestions

Chris had no idea that the system was guessing my language, this was not at all their intention, which made me wonder if this was the doing of FamilySearch instead.

@Helmut: Sorry if we are hijacking your question. This new location feature just raises so many questions that it is hard to find the best form for a discussion.
Eva, Above the birth date field is a new box for Language. If you set that to the language for that profile then the suggestions should come up in the proper language. The biggest problem I have found is when people immigrated, I have one where the birth location should be in German and the death location should be in English, but I am sure that I can work around that.

I believe that WikiTree just links to the FamilySearch database to get the suggestions and does not store anything that you do not enter on WikiTree yourself, but I could be wrong.

Well, Dale. According to Chris' response in the other thread the system should not be guessing that my language is Swedish, so there should be no need to use the Language box to compensate.

What I was interested in here was in second-guessing how the system works - because whichever profile I test, the system does guess where I'm on the Internet from, which it should not be doing.

Eva, If you set the language for the profile it should give you results in the language that would be correct for that profile.

My theory I dont know if this is the database we use I just guess

I think the problem is the presentation of the location. If you search on Hulu at https://familysearch.org/stdfinder/StdPlaceLookupResults.jsp 
you get an insane number of Hulu Jönköping see pic below

I guess the column Geocode is the GPS coordinates and if I place them on a map it's far from the same place ==> 

  1. we need to store the ID and GPS coordinates in WikiTree
  2. Just text searching feels impossible..... to understand what Hulu you find
     

See my Google map with some of the locations for Hulu


Big pic 



Big pic

Big pic

The only thing we can trust I feel is the identification numbers in the column to the right and the GPS coordinates

The rest is just Hulu Hulu Hulu and a small taste of how complex locations are... As an addictive iceskater I know the feeling when they tell you its ice on lake xxxx in the old days and you travel by car 100km just to realize that you are at another lake with the same name and no ice ;-)  Now we have dedicated iceskating web with GPS coordinates, GPS tracks, reports and online maps and possibility to speak in reports when iceskating how good the ice is or not....

 

 

EDIT I checked the word hulu and it is old swedish for a small bad place to live so that is maybe the reason we find hulu at more locations. 

Found another place that looks like the interface we should have

https://familysearch.org/research/places/results?

A picture explaining what we have and what would be the best


Big pic

URL Family search map 1669312 Hulu
 

Nota bene this interface also have spatial search ==> find all places 5 miles from this location URL feels magic compared to what we have today

 

  • Feels next search option would be Find Profiles in Wikitree connected to this area and 5 miles away between 1700-1800 ;-)  Like working with today's Categories but version 2.0 
     
  • The History Pin search is also really cool showing old houses from the area

2 Answers

+3 votes
Unlike the place name in the profile's data field, which should be in the language and reflect the time of the person profiled, categories are in specific language streams.  The English language stream is best developed, but if you click on "find" at the top right coner and then "categories" you'll see actual or potential streams in almost two dozen languages.  So I would expect to see Aachen in both the English and German language streams.

A second principle which distinguishes us mightily from some other internet based genealogy sites is that we try to be historically accurate in our place names.  So the question really is, what did the people of Aachen, and its ruler, see themselves part of in the year 800?  While we refer to Charlemagne as the first Holy Roman Emperor, that specific term was not used until the 13th century.  Certainly, Aachen was in neither "Germany" nor "Deutschland".  Frankish Empire, perhaps?

Sorting this out may seem like a lot of trouble, but there are some benefits -- categories try to group like profiles together for genealogical benefit.  If a category was siimply for Aachen, Germany, you would have Charlemagne grouped together with the woman who runs the bakery down the Aachen street in 2017.   By giving Aachen a category name specific to a particular time period, one is not only historically accurate, but will group together others who were present in that place in that time period.
by Jack Day G2G6 Pilot (462k points)
edited by Jack Day
+5 votes
Let me add another note -- the name data base from Family Search has been presented to us as an option which may provide assistance to users of WikiTree.  Extensive discussion has already established that in many instances it does not provide either an option nor assistance.  

WikiTree users in collaboration with others in geographic-based projects and on G2G already have procedures for determining what makes the most sense TO US with regarding place names at specific periods.

The danger in employing a database like Family Search's  lies precisely in giving the false impression that (1) Family Search's names are always correct and reliable and (2) that they should supercede the names that WikiTree users have determined appropriate through collaboration and discussion.  The use of the word "standardization" in the query itself leads me to worry that this process of subverting WikiTree's own integrity has already begun, and that there is now a policy that the Family Search names are to be used in preference to those WikiTree members have taken the time to consider and develop.

As an advocate that Marylanders born before 1776 were born in the Province of Maryland, and that Indonesians born before 1945 were born in Hindia Belanda (the Indonesian language equivalent of Dutch Indies) I very strongly oppose the notion that we have given a standardization responsibility over to Family Search.
by Jack Day G2G6 Pilot (462k points)

Jack, I meant "standardization" to reflect the need for us on WikiTree to do something in response to the inadequate FamilySearch database. I fully agree with your concerns about that database becoming nolens volens the "gold standard," after all, there have already been instances where correct locations were changed to the database suggestions.

Perhaps we do need to put up some extra effort to do that. A first step could be to ask Aleš Trtnik to produce a list with all locations. I'm sure the list would be immense but one could start whittling it down by flagging the obviously false: for instance the use of English names for non-English speaking locations. It could then become an objective for regionally based projects to get the names for their locations right. During this process we could flag other obviously false locations such as Aachen, Aachen, Rhein, Preußen, Deutschland for 9th century Francia ...

But I suppose institutional inertia will make us keep the FamilySearch database instead.

Related questions

+7 votes
1 answer
220 views asked Sep 10, 2018 in WikiTree Tech by Pip Sheppard G2G Astronaut (2.7m points)
+10 votes
4 answers
+25 votes
11 answers
+16 votes
3 answers
+8 votes
0 answers
+16 votes
6 answers
532 views asked Dec 18, 2017 in Policy and Style by Living Anonymous G2G6 Mach 5 (51.6k points)
+11 votes
4 answers
+12 votes
2 answers
+6 votes
1 answer
273 views asked May 6, 2023 in Policy and Style by Katrina Lawson G2G6 Mach 4 (48.6k points)

WikiTree  ~  About  ~  Help Help  ~  Search Person Search  ~  Surname:

disclaimer - terms - copyright

...