Why is 511 checking middle names when the definition says it checks first names?

+9 votes
386 views
The 511 db check seems to be checking the middle name field as well as the first name field - causing countless "errors" in British names, where it is common practice for the mother's maiden name (LNAB) to be used as a middle name.

I suggest that this check is changed to check the first name field only.
in WikiTree Tech by John Elkin G2G6 Mach 7 (70.8k points)
retagged by Dorothy Barry
Perhaps it is because in the profiles created by GEDCOM loads the first name field contains the middle names as well. There are, apparently, a number of problems in identifying what is a middle name on a GEDCOM and what is a multi-name first name and the conclusion was to include all names in the first name field.

What John said is correct. The great majority of the 511 “unique name” db errors are resulting from Middle Names (usually a traditional family name or maiden name of mother, grandmother, etc.)

In the past 2 weeks, I have found 99 False Errors attributed to profiles that I manage, based solely on that “unique” middle name - all properly in the middle name data field. 

What a needless waste of time!

Just to better understand what you think should be tested and why/why not and do you have an example?!?!?

  1. What is the bad/good that the middle name is tested?

    causing countless "errors" in British names, where it is common practice for the mother's maiden name (LNAB) to be used as a middle name
     
  2. When you check the Middle name with problem at ShowFirstNames.htm how many other profiles do you find with this name in the field..... - could be an indication that we use the field different is the lesson I learned about Swedish Patronymic...
     
  3. Today in Wikitree you can add the following names. What do you think should be included in the First Name test?
     
    1. Prefix
    2. Proper First Name
    3. Prefered First Name
    4. Middle_Name

      This should be the full middle name(s).

      If you only know a middle initial, use that in its place and mark the status as "uncertain."

      If you don't know the middle name, leave it blank.

      If you know that the person has no middle name, leave the field blank but make it clear to others that it's not just unknown by marking the status as "no middle name."

    5. Other Nicknames
       
    6. Last Name at Birth
    7. Current Last Name
    8. Other_Last_Names
    9. Suffix

      You also have the following subjects that could impact what users add to different fields in WikiTree
       
      1. Name Displays
         

 

Question: Maybe we should define another error that check the content in the Middle name? 

 

 

Why should we care?  As long as people do what makes sense for their ancestor, we don't need the automated error checking to intervene.  There are plenty of cases where the software helps; this is one that is mostly better for humans to find.

Re Mikey what do you argue about

  1. its stupid with Project:Database_Errors 
  2. that we shouldn't try to have a name standard at all i.e.
    1. Names in what fields
    2. Locations standard
Reason I think that we should care and at least try to do it in the same way or second best understand that we doesnt do it the same way ;-) see (comment Patricia) is 
  1. that matching/finding duplicates is so much easier if we have a name "standard"
  2. Easier for other researcher to understand that this was the name at birth
  3. If we start matching WikiTree with e.g. Family Search then if we have some consistency it will be easier to find matches in the fields they have (see below) 
  4. it's much easier to find mistakes with something like Project:Database_Errors if we have some consensus were e.g. we add the patronymic. If we start adding patronymics also in the first name fields ==> we can never find a mistake that someone has done it the "wrong" way....
WikiTree is today rather unstructured we have some name fields and one location for birth and one for dead compare Family Search who has templates like Cyrillic, Japanese, Khmer, Korean..... just because this is rather complex and you gain benefits if you do genealogy and use a database with specific fields for different purpose.....
 
 
Family Search validates locations ==> they have control and can find everyone born at a specific location. In WikiTree today we have no standard and no GIS spatial searching ==> difficult to find an ancestry event not far from my current position....


Hi Magnus,

In response to your question about which fields I would expect to be included in a "first name test", I think that's obvious - the only relevant fields are the "proper first name" and the "preferred name", as these are the only fields that contain first names.

We could also have similar tests for some of the other fields:

  • elements of the "middle name" field could be validated against a combined list of all distinct "first names", LNABs and CLNs
  • the "prefix" field could be validated against commonly occuring entries in the "prefix" field, and also checked for length (the Euroaristo Project naming conventions say this field should be short)

Additionally, we could check that the "proper first name" field contains only one element - although we would then need to be clear on the handing of commonly occurring compound first names (such as "Mary Ann" in English speaking countries - my understanding is that this should be set out as: "Proper First Name"="Mary"; "Middle Name"="Ann"; "Preferred name"="Mary Ann").

For an example of the English custom of passing on the mother's maiden name (LNAB) as a middle name of her children, take a look at the Orlebar faimily - such as Augustus Scobell Orlebar.

Thanks good answer John Elkin

I start thinking that we need to look into different scenarios and what is best practise and alternatives. Below what I have found and also some good pages e.g.

  1. English custom
  2. Name Fields for European Aristocrats
  3. Cyrillic names
  4. Scandinavian Patronymics
  5. Name Fields for Czech Names G2G
  6. Dutch Roots Naming Convention
  7. New Netherland Settlers - Naming Convention
  8. Welsh Aristocrats
Got the feeling from Mr Database Error project Aleš that he was open to introduce "new" errors that could check that we following naming guidelines..... ==> we need to explain the rules so a computer understands.... 
 
Now I feel I think we are on a level that its difficult for a tired Swedish old man has problem to understand ;-) what is correct ;-)

Best would I think to define rules like
1) Person born in Sweden
2) Before 1900
3) Father name is Anders, Carl, Karl, Olof, Olof......
4) Then if we have a Andersson name or a Anderson name it should be in field xxx ---> else is marked error 2012 

3 Answers

0 votes
The genealogy program I use most only has Given Name, not first and middle.  That not only works, it's actually better.  And to improve searching and matching looking at dates and location helps narrow it down.  WikiTree needs to add location to the search function so that we're fishing in the right pond.
by M Anonymous G2G6 Mach 4 (47.1k points)

Why is it better?

I understand it's easier but genealogy is not easy.....

 

Maybe skip the name field and just do as below ;-)

It's better because (A) we wouldn't be having this discussion, and (B) we wouldn't waste everyone's time on stupid control issues like splitting up Mary Ann and putting Ann in the middle name field.  If the software can find real (i.e., obvious) errors, the rest can and should be left for people.

Ok problem is that a software today need machine readable data and rules ==> we need to have some structures... ce la vie...

Maybe life will be easier in the future...

+3 votes
For now I only get First name in database dump. I think the First name I get consists of Prefix + Proper First Name: + middle name. Other name fields are skipped, so prefered name and other nicknames fields are skipped. I think this will change in future, but for now it is so. So I cannot distinguish between first name and middle names. That means that error 511 is best I can do with this.

The confusion is because field is named First name but it is actually Prefix + Proper First Name: + middle name.
by Aleš Trtnik G2G6 Pilot (478k points)
I think you should leave Preferred Name out permanently, or just do minimal checking for really obvious errors.  That field is for the users to enter how we want the name of our ancestors to be displayed.  Mostly the same applies to nicknames.
+2 votes
If someone is born in the Netherlands every middle name is a hard error, because the dutch law don´t know the term middle name.

It could be me, but middle names are only used in a very small part of the world, so maybe there should be a check on a location and if it is in that very small part of the world the middle name should be a hard error. Personally I don´t want to see the middle name, because to many people use it and I spend too many hours correcting and moving first names. If you see Jeroen Willem Van Van Dijk. Then that is an merging error. But Mary Ann Ann Smith is also very likely a mistake.
by Jeroen Mathijs Willem van Dijk G2G5 (5.3k points)
edited by Jeroen Mathijs Willem van Dijk
I can add new error for that. Also for Slovenia. We don't have term middle name. Probably some other European countries. Local people please add countries here, that don't and didn't have middle name.

But if the person moved to US, than he might have middle name? Or not?
I was born in Holland and given two names.  In the US I have one first name and one middle name.

Related questions

+24 votes
7 answers
+7 votes
1 answer
177 views asked Jan 5, 2017 in WikiTree Tech by Brett Rutherford G2G6 Pilot (120k points)
+10 votes
5 answers
+6 votes
0 answers
103 views asked Feb 28, 2017 in WikiTree Tech by Mary Cole G2G6 Pilot (101k points)
+6 votes
2 answers
+16 votes
1 answer
+13 votes
4 answers
+10 votes
1 answer
97 views asked May 15, 2017 in Policy and Style by Eva Ekeblad G2G6 Pilot (371k points)
+12 votes
1 answer

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

disclaimer - terms - copyright

...