:( Why are all of the 511 Unique names, that are False Errors back again? Do we have to delete them every week?

+14 votes
801 views
Half of my ancestors are Norwegian and they all come up as 511 Unique Names on the error report, so I spent a great deal of time doing False Error for each one, so they would go away and I could focus on the real errors, but now they are back on the report :(  Do we have to delete them every week?
in WikiTree Tech by Patricia Roche G2G6 Pilot (779k points)
retagged by Dorothy Barry
Same situation with German names :(
Also dutch,

http://www.wikitree.com/wiki/De_Volder-1

most of the errors in the list are 511, in old Dutch records are a lot of different spellings for the same name. In the profiles you write down the names as in the birth certificates etc is written.

I suggest to remove these type of errors for first and Lastnames.

No please let it be there I think they are very useful. It's a way to find errors....

As the function is based on statistics in the WikiTree database we will get less and less "false" errors see DBE_511

http://www.sdms.si:92/wikitree/ShowFirstNames.htm

Did a search on Matje
 

Total: 19 FirstNames for "Matje".

SHOWING FIRST 4

First Name Total Public Gender Empty Gender Female Female (%) Gender Male Male (%) Name Type
Matje 42 39 0 38 97.44 1 2.56 Female probably
Lena Matje 1 1 0 1 100.00 0 0.00 Undefined
Martha / MATJE 1 1 0 1 100.00 0 0.00

I suggest to remove these type of errors for first and Lastnames.

LastNames are not yet checked.

Comelisse is unique in Wikitree....

Total: 1 FirstNames for "Comelisse".

First Name Total Public Gender Empty Gender Female Female (%) Gender Male Male (%) Name Type
Matje Comelisse 1 1 0 1 100.00 0 0.00 Undefined

 

I see Comnellise is wrong it has to be Cornellise. Sorry wrong example.

 

Help section for Validation updated please see link

2 Answers

+7 votes
Can you give me some wikitreeID of such profiles.
by Aleš Trtnik G2G6 Pilot (796k points)

Joop, profile De_Volder-1 was marked as false error and temp hidden (you don't need to use both. False error is enough) today, so the error wasn't marked as an error before.

As you can check here http://www.sdms.si:92/wikitree/ShowFirstNames.htmComelisse is very unique. Your profile is the only one with this name. So if spelling is correct, it was ok, to mark it as false error.

OK, thanks for your comment.

From the 256 errors for Zeeland there are approximately 130 errors with the code 511. Most of these names are correct spelled so if I marked these as false error the are next Monday disappeared?
They will disappear immediately, if you refresh the error page. But you should not just mark them all as false error. Only those that are correct. I would also like feedback on how many were misspelled and how many were correct. I cannot really estimate the quality of these error.

So please When you will go through 130 errors (511), make note how many were correct and report back.
0ne of the problems with checking this is that a lot of these profiles are pre 1500 and Unsourced or not properly sourced. So it's hard to check the correct spelling, if there is one, of this people from the middle ages.

What to do if we  simply don't no the correct spelling?

Beside that, this is a very good tool you have given us.for checking errors,  my compliments Aleš
@Aleš,

I checked 50 of the 130  511 errors (time period 1750 -1950).

False error: 22

Correct error: 24

No sources found: 4

So approximately 50% of these errors are correct.

Names of ancestors before 1800 were often written in different ways depending on the public official. In addition, they are often difficult to read in the old books, so difficult to check.
50% is quite ok. I was afraid it was lower. Over time this ratio will increase as database accuracy increase. Also filtering out -son... will improve that ratio. I might also add learning capabilities from reported false errors, although there are cases, where people just clicked false error instead of correcting it.
+10 votes
It is true, that last week error 511 was checking names only for dead people. Now it checks name for all profiles. Maybe that is the reason, you think errors reappeared, but they are different ones. I also reduced needed frequency for a name to be ok from 20 to 5, so there are fewer false errors this week.
by Aleš Trtnik G2G6 Pilot (796k points)
edited by Aleš Trtnik

Here is the report for all names in question. 

http://www.softdata.si/osebe_staro/ales/wikitree/FirstNameSon.htm

So swedish endings .sen, .son, .datter, .dotter, .dtr should be in last name at birth? Maybe an error for that?

Yes if we agree. 

I will create a new topic later today regarding handling patronymics in Wikitree. 

As said above in Sweden I guess most people agree that Andersson is not a First name

In Norway I feel they have enormous complex names e.g.

we need to check with Norwegian people


A rule could be 

  1. If person is from Sweden or grandparents
    1. Sverige is Sweden in swedish
       
  2. check suffix 
    1. son or sson if the person is male 
    2. dotter - if the person is female
    3. maybe also check if
      1. we have a father with the name
         
  3. In Sweden in the old days (before 1900) you took the father's name and added daughter or son. After 1900 you started to "inherit" the father's Andersson name without having a father called Anders but someone before had the name ( I assume the same way  

 

Ok, I am away until monday, so I will look on then if agreement was achieved.
For Netherlands (Nederland) I guess the most common are "-sz" and "-dr"

So swedish endings .sen, .son, .datter, .dotter, .dtr should be in last name at birth? Maybe an error for that?

Please NO - There are opposing understandings and use of Scandinavian names on WikiTree, and to enforce one, will hamper research for the other. In rural farm areas their names were NOT as we understand them: first-middle-last.  They were a statement about that person - their title. At home they were addressed by just 1st name, but when spoken of or to, out in the community and in the records, they were ____ son/daughter of_____ living at_______. 

There could be thousands of ____son/daughter of______, but they are _____son/daughter of______living at______ of which there is one (or few). Their farm name is critical in research/identification of who is who.  On a land contract, the person was ____son of _____ living at xxx  at the top of the contract but signed the bottom ____son of _____ living at ooo,  as his title changed when he became legal owner of new farm.  So for us that use the old Bygdebok (farmbook) records for genealogy research, we put 1st name and who their father was in 1st name field and their appropriate farm name in last name field. If they moved the LNAB is their title at birth and Current is their title at death, with any other in-between titles in other names field.  It is critical for us Bygdebok (farmbook) researchers/genealogists, that search results bring up the statement ____son/daughter of____lived at_____born_____, as they were recorded, or our ancestor looses their identity/tradition and we would not know one Ole Olsen for the million others. 

Some folks prefer not to use this naming tradition and that works for them, but for those of us that follow this tradition, our tree branches will be useless if we put who their father was in LNAB or even in Middle name field. Google and WikiTree search results need to show their title correctly: ____son/daughter of____lived at____born_____.

 

Wikitree help on the subject Last_Name_at_Birth

It is generally a family name but it could be a patronymic or whatever other standard is conventional for the person's time and place.

Report 2016-05-29 Error 511 - Unique firstname (spelling)

  • In total 293 826 errors 25% of all errors reported
    • 0000-0000 number errors 52 188 
    • 0001-1499 number errors  6 740
    • 1500-1699 number errors 13 072  
    • 1700-1799 number errors 23 234
    • 1800-1899 number errors 118 191
    • 1900-1999 number errors 78 882
    • 2000-Now number errors 1 570
    • Now-9999 9 

We also discuss this in the Swedish Facebook Wikitree group please join

I see a suggestion that seems to be saying that a programmer will be moving names from one field to another.  Please do not use a program to change my data!  Do use these rules that you are developing to generate error messages that tell us what the program thinks is wrong.  Do not attempt to "fix" the data that you find!

There are just too many possible ways to go wrong with data like names and places.  We have names from many different countries with many different customs.  We also are quoting from records which contain many different versions of the same name; the farther back we go, the more this is true.

Thank you for your efforts to correct the database without making it worse.
I don't think any automation will be done in near future. WikiTree staff is very conservative about that, so don't worry. Although some errors could be corrected automatically by Chris. He could remove text Unknown from location field for all of them at once, but...

Re Pat 

Interesting to look at this WikiTree polll that most people think Bold is the ok see link ==> not a big difference compared to use a bot ;-)

If we look at Wikipedia they mark articles in the Talk section if an article can/should be updated with a bot. I think WikiTree lack people like  Aleš who could do bots so its a non issue...

 

 

Related questions

+25 votes
7 answers
+7 votes
1 answer
250 views asked Jan 5, 2017 in WikiTree Tech by Brett Rutherford G2G6 Pilot (127k points)
+11 votes
5 answers
+6 votes
0 answers
194 views asked Feb 28, 2017 in WikiTree Tech by Mary Cole G2G6 Pilot (109k points)
+7 votes
2 answers
+16 votes
1 answer
+12 votes
1 answer

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

disclaimer - terms - copyright

...