We depend upon records to reveal the “truth” about our pasts.
Yet sometimes records have anomalies.
Some are amusing or humorous.
Some are interesting or weird.
Some are peculiar or suspicious.
Some are infuriating, even downright laughable.
Yes, “Records Say the Darnedest Things.”
Records Say the Darnedest Things: Darned Gender-Specific Names
With the public slowly getting access to the New FamilySearch Tree, it bears repeating lessons learned.
When I first got access, I went in and started cleaning up the riff raff. The more ancestry you have in the Church of Jesus Christ of Latter-day Saints, the more garbage you will find. For five generations Mormon genealogies were hand copied (or typed) with mistakes (and typos) from parent to child among an ever widening fan of cousins. Reconciling the many different versions is herculean, but to guarantee it would be done wrong, FamilySearch produced the Ancestral File using error-prone machine combinations of patron submissions. Then they provided GEDCOM download and PAF’s hideous merge feature. Lest all this hard-earned garbage die too quickly, FamilySearch preloaded it into the tree and dared participants to clean it up.
That brings me back to my first session in the tree. I only meant to poke around and see what was there. But the errors were obvious (and legion). The temptation was too strong. I started correcting. Just small stuff at first. Several hours passed without notice and I realized I had corrected dozens of errors by sight—without consulting my sources.
Lesson learned? Don’t do that!
FamilySearch Chief Genealogical Officer, David Rencher, suffered the raw end of someone making such a change. The tree says his father’s name is Joy Thomas Rencher even though his grandfather’s name is correctly shown as Thomas Jay Rencher.
Some kind soul “fixed” this, changing the son’s name to match his father’s. Like me in my first foray into the new FamilySearch Tree, the person made changes without specifying a source. Had they consulted sources, they would have found Joy’s death certificate:
David Rencher confirmed that his father’s name was Joy. He said that Joy is a common male given name among the Danes in the White Mountain area of Northern Arizona in Mormon Settlements. He knows of six men named Joy in that region.
Yes, records say the darnedest things. But they should be consulted prior to making changes in the New FamilySearch Tree.
Gender specific names are just one aspect of the larger issue of inaccurate poorly sourced additions and "corrections".
I was just talking about such tree issues with my co-worker last weekend on my weekend to volunteer at my center. He is an IT professional and our conversations re FamilySearch often revolve around programming and technical issues (we're not fans of Java web apps in general or the way the programmers seem to drive the site instead of user requirements (sorry!)).
But that is implementation/execution. The overall requirements of the project of course have a spiritual basis, which I think I understand well despite being non-LDS. So the decision has been made to make the tree accessible and easy to use for members (and soon non-members as you point out) at some expense of accuracy. But the question is how much inaccuracy can be tolerated? That seems not to have been specified from anything I have heard or read. Indeed it seems open-ended.
So you have people uploading junk gedcoms, providing no or poor sourcing for making the type of "helpful" corrections you talk about. And it is leading to one big snowball of error that will gather momentum and grow exponentially once the general public can use the tree.
As a volunteer my director was able to get my FS account tree-enabled early before the mass of non-LDS are allowed to do so. So naturally I navigated around a bit and entered the names of some ancestors born late 1700s who have lots of descendents. And of course there are lots of dupes. So I thought why not correct a couple. But the dupes vary widely in asserted parents, vital dates and places. So the merges would be very complicated and time-consuming. And that is on top of the fact that it is all a big game of wack-a-mole since new erroneous gedcoms will just restore the status quo ante in no time at all. This is not an incentive to put in time correcting and I indeed am not going to do so.
It would be nice IMO if the leadership could put some reasonable limits in place to make the FS Tree better than the Rootsweb WorldConnect and Acom trees with all their dupes and errors. Such steps could possibly be:
1) Stop allowing gedcoms, at least for the US - it is highly unlikely a very large US gedcom has many truly new unduplicated people (past the 100 year limit), and the process of hand typing new persons allows time and thought to accuracy and citations.
2) Set a cutoff date of say 1900 or something, before which records absolutely cannot be "corrected" without legit sources, i.e. not "common knowledge" or "my grandma said".
3) Make the system more robust in terms of trying to prevent dupes from being added even by hand.
Obviously the leadership of the Church would not want to see a spiritual detriment, as in members feeling that ordinary members cannot participate and be able to contribute to temple ordinances. But surely such spiritual purposes requires a commitment to accuracy, even if it is more time-consuming. It is a balance of course, but as of now it seems to me that there is a brick on the side of the scale for widespread accessibility and a pebble on the side for accuracy.
AI, thanks for your continuing items pointing toward need of evidence for stating genealogical conclusions.ReplyDelete
Mike F., you expressed reluctance to spend energy on corrections in the NFS tree, " . . . since new erroneous gedcoms will just restore the status quo ante in no time at all."
You may be happy to hear that it is planned not to allow GEDCOM uploads in the future. According to Ron Tanner, there will be an interface whereby patrons can upload a GEDCOM into Pedigree Resource File, then selectively add to nFS-Tree where persons are not duplicated. He stated this somewhere in the getsatisfaction.com message board, but I could not quickly find it.
There is a difference between nFS-Tree and "the Rootsweb WorldConnect and Acom trees with all their dupes and errors." The latter two are composed of myriad individual trees (although Ancestry.com created a horrid tree mess by having a computer program combine many thousands of trees that existed around 6 years ago in a format that it no longer supports - producing the notorious OneWorldTree, which is still available to copy from, ecccchhh).
The individual trees on those sites and many other internet sites will necessarily contain duplications and be error-riddled. Many of them were derived from material found in IGI, Ancestral File and Pedigree Resource File, as well as other non-evidentiary sources, although there are sterling exceptions built by good researcher-genealogists.
Many with current access to the nFS-Tree either do not understand that it is a single tree, or object to not having their own untouchable tree. A great deal of genealogical education is called for, and at present the volunteers are pretty much the front line for pointing to the wealth of resources available on the FS site -- though there seem not to be ready links to the wiki and other parts of the Learning Center from nFS.
Even implementing your recommendations will not soon fix the present problems, sad to say. Since one cannot rely on a computer program to discern what "legit sources" might be, or what within them might be applicable evidence, you are proposing labor-intensive work on the part of thousands.
How does one ask Familysearch.org to correct a Pedigree Resource File?ReplyDelete
A copy of an obit has been posted to show the death of Henry Farley Byerly for a pedigree under his name.
The obit is from a Clinton, DeWitt County, Illinois newspaper and records HFB's death occurring in Kenney, DeWitt County, Illinois. The Pedigree Resource File shows HFB's death in Austin County, Texas. Texas Cemetery is in DeWitt County, Illinois.