News on Database errors project (31 July 2016)

Question

Analysis was done on data from July 31th 2016.

Detailed statistics are available on http://wikitree.sdms.si/default.htm in Statistics section.

Comments

Unfortunately in my family Don and Donna are very common first names in my family and they are all errors, (Prefix in first name)....I understand marking them with False Error, but, it takes time.....

---

And why is Reverand an error as a prefix?

---

Isn't Reverend correct spelling? I am not that sure in my spelling, but on Google Reverend in used much more.

---

True. I also heard for both names. I will allow them as names. I will recalculate 73x and 74x errors after some updates.

---

Is Dona also a common name? Or is it misspelled?

---

Dona is not common, only Donna..and you are correct about the spelling on Reverend....sorry....

---

73X and 74X updated.

---

Dona is not common in English, but I think it's common as a name spelling in Spanish.

---

Dona is now only prohibited in suffix. It will be updated on next update.

Answer 1

I probably missed this in all the conversations that have been posted, but why does "unique spelling of first name" also include what is listed as nicknames and middle name?  Of course, it will come up as an error if all those names are included in "first" name.

Also, if the unique spelling of the first name is all that can be found in sources, does one click on "false error?"   Will that remove it from the error list?

Comments

I'd say it depends on the quality of the source.  If you're relying on what other people of unknown ability to spell or write have produced, how can you assume a strange spelling can't be an error?  And lots of times something an earlier genealogist found on a document is then copied by many others, so the frequency doesn't help.  If something is found on a marriage certificate and a death certificate then you can probably assume it's not an error and click on false error.

---

This error is from old database dump, where first name was joined all names together. In new dump we had prefix, First name, proper name,... and I tried to make the error the same as it was, so it is combination of all names. 

But this error is checked by words and one of the words is very rare and is probably a typo. If you click link on name, you will get usage statistics of each word. If you think, the name is spelled correctly, click False error and it will disappear from errors. In prepared lists it will be removed on next dump (each monday).

 

Answer 2

Updated errors 607, 637, 667 Location spelling

Added spelling verification for a few locations. I will add more of them in time. For now I check Massachusetts, Leicestershire, London and England. I added all words, that appears more then 10000 times and are longer then 9 letters. Shorter words have more exceptions so I will add them on request. Please check error list for actual locations. I am sure I missed some.

You can find correct spelling of checked words here. There you can also add new locations to be checked. Each added word must be checked for actual locations. London has exceptions for Lyndon, Loudon, Loddon, Longdon because they are actually places.

Comments

Londen should also be an exception. That's a Dutch spelling for London.

---

Aren't we supposed to use the spelling the people who lived there would have used? Why would someone referring to London need to spell it as Londen?

---

Many Dutch people migrated to England in the 1500s, and there were Dutch churches there. For an example, see Der_Kinderen-5. That profile uses English spelling, but if it had been written by a Dutch person, it would have had spellings like "Londen." 

---

I don't see that as a good idea. Then we would have to add each city in all 300 languages, which is pointless. Guidelines are quite clear on that. Local or english version of  the name. I checked and Londen is used cca 60 times. I did exclude Fondon (Spain).

---

Fair enough, but if names spelled in other languages are going to be displayed as errors, users should have the ability to mark those errors as "false errors." (The most I could do with the instances of "Londen" that I saw in the error report was "ignore for 30 days.")

---

I can add false error, but I wouldn't like to. Most of the errors must be corrected, to be able to put them on the map. I can add Londen to exceptions, but it will not be recognized as London and will be again pointed out as an error in future, as new location validation will be done.

---

Here you can see all profiles, that uses Londen.

http://wikitree.sdms.si/function/WTWebProfileSearch/Profiles.htm?Query=Londen

There is also New Londen, CT. Is that also dutch spelling?

---

The place name that caused me to comment about the spelling "Londen" was "Oos Londen" (I saw it in some entries in a place-name spelling errors report).  From the report you linked to above, it's apparent that "Oos-Londen" or "Oos Londen" is a place in the Cape Colony or Cape Province (South Africa). Philip van der Walt should be able to confirm that.

New London, Connecticut, is not spelled New Londen. (There was a Dutch settlement in the vicinity of New London, but that settlement didn't name itself for an English city.)

Almost all of the instances of "Londen" on that report you linked to are Dutch spellings, including people's names (like Van Londen) and that place in the Cape Colony.

---

Since Oos Londen is valid city, I will add Londen in exceptions. 

---

Thanks.

Answer 3

Custom checking and correcting

You can check spelling of any word on http://wikitree.sdms.si/default.htm in group Analyse item Location spelling. There you can also manually check any location spelling and view misspelled profiles and correct them.