Shared publicly  - blindly copying FamilySearch collections
#genealogy #Ancestry #hot #warning #FamilySearch
After two hyperbolic press releases extolling agreements between and FamilySearch, we are experiencing what the agreements really mean: is blindly copying FamilySearch databases, regardless of quality, and not even performing sanity checks.

Dutch genealogist +Yvette Hoitink noticed that Ancestry had added a few collections from FamilySearch, purportedly containing a few million Dutch records, to wit:
■ Netherlands, Select Births and Baptisms, 1564-1910 (in Dutch) [6.2 million records]
■ Netherlands, Select Marriages, 1565-1892 (in Dutch) [1.4 million records]
■ Netherlands, Select Deaths and Burials, 1668-1945 (in Dutch) [0.6 million records]

Her blog post Major problems with new Dutch records on enumerates all the obvious issues with these collections and then some, including an issue few people but Dutch speakers would notice: "Burgelijke Stand" isn't a placename, it is Dutch for "Civil Registration".

Now, her blog post discusses the issues with these databases on, but it is not very likely that these issues are unique to, as a copied database should be identical to the original. Sure, errors may be introduced by an erroneous copy process, but I did a quick check, and it does seem the issues originate with the FamilySearch databases:

The image below is the same record used as an example in Yvonne's blog  post, but the FamilySearch original instead of the copy. According to FamilySearch, Gerrit Hendrik Aal was born in "Burgelijke Stand"...

That issues originate with FamilySearch does not mean doesn't deserve any blame. Subscribers who pay hundreds of dollars per year should be able rely on some minimum level of quality control.
John Tierney's profile photoTony Proctor's profile photoTamura Jones's profile photoYvette Hoitink's profile photo
Quality control from  Not likely, especially with databases they license from others.
On a similar note, I sent Ancestry a message recently about an issue with their data connection to the Ireland 1901 & 1911 censuses where the place names all come in the form Townland, DED, County.

An example from my family looks like this:
"Creggan, Moyclare, Kings Co."

But, Ancestry interprets that location as in a place called "Kings, Colorado, Ireland", mistaking the Co. for county as the abbreviation used for the US State.

Very odd. Have not heard back regarding my inquiry and see the issue persists.

Thank you for picking up this story!
I agree that most of these errors already existed on Familysearch and that they are probably the source of most of the errors in current Ancestry Member Trees. But on Ancestry it is much easier to attach these sources to the trees, where information will be copied into the tree unless you edit it. On Familysearch, that was a manual action. I predict that the 'Burgerlijke Stand' error will multiply much more quickly now it is on Ancestry. 
+Yvette Hoitink,  I am aware of the horrible automatic place name (mis)matches you mention. I have seen these mentioned before.
I do agree that's shaky leaves do not only help users find connections, but in practice also lead to copying of data, good or bad - so I fear your prediction is accurate.
Whether FamilySearch is the source of most errors in Member Trees I dare not say, but it would not surprise me.
Regarding your infographic, I would not classify the "Burgelijke Stand"  as a software error, but as an indexing error.
In my comment to +Bob Coret at I discussed the causes I suspect for these problems. I didn't put that in the infographic for two reasons: 1) 'How software, database and indexing errors corrupt our trees' would be a bit of a mouthful and 2) infographics tend to lead a life of their own so I didn't want to put anything in there that I might have to change later on. So I added the explanation in text, that I can always edit if I discover new evidence. 
Add a comment...