Data quality for collaborators

Started by Erica Howton on Saturday, July 10, 2010


Showing all 8 posts
Private User
7/10/2010 at 7:37 PM

Interesting idea, Erica.

7/10/2010 at 7:45 PM

I would like to have the ability to prevent future merges from screwing up data after it has been sorted, but I'm not sure it is really possible. Perhaps if staff is not willing to implement such a lock, it might be possible to create a script to help quickly set things back the way they are supposed to be. This would sort of be like the scripts that monitor wikipedia entries and quickly undo defacements.

7/10/2010 at 7:50 PM

i understand your frustration!
but how would we evaluate a person as being an "expert" or whatever???

Erica, interesting discussion here.
A number of minor comments before I get to the real beef.

When people tell you to fall back on the "primary manager", they generally mean, NOT to go ask them, but to select the data entered in that column of the resolve data screen. The logic behind this is that the "main" profile has probably been reviewed far more often than some single manager copy. Of course, being been merged the most often can also mean the opposite - that the data has been trashed oftenest. In MANY cases, asking the "primary manager" a question WILL produce the best results, because over time, people tend to become managers of the profiles they are most interested in. For example having spent much of two years merging the biblical tree, I now manage large parts of it, despite not having entered these profiles.

In this regard Bjorn Brox is actually somewhat of an exception, he manages some 250,000!! profiles, more than ANY other Geni user (IIRC), most of which were transferred to him by Geni staff, from abandoned accounts. This was done for a number of reasons, primarily because he's so helpful and knowledgeable AND has one of the largest sets of Collaborators and added "relatives". So giving him the profiles, makes them most accessible. So in most cases, not only did he not enter the profile, he might not even know it exists.

With the present system, there IS a very good chance that Mimi's work on the Plantagenets will be undone, by users who are less careful. This is PRECISELY why the Curators concept was proposed. Geni HAS acknowledged to a limited extent, that they probably will need to implement just such a system. As to a time-line for implementation? Nothing definite. There are a number of much higher priority issues that Geni is presently working on, such as revisioning profiles and being able to undo changes.

A clean-up script, that is an automated process, would never be smart enough to keep track of such things. This is why we need Curators, who could keep track of things AND be the single authority able to actually change these curated profiles.

The problem with extending this idea to all public profiles, is that 1) there are WAY too many of them (millions), Noah of Geni thought he'd need hundreds of Curators. Those who made the proposal think Geni will be hard pressed to find a couple dozen... 2) at least for the less famous, no real way to "validate" the data. 3) No "public interest" regarding most of these profiles, so no real need to get involved in people's "personal" trees.

I think your idea IS worth considering. It would take a much larger secondary group in addition to the Curators. I'm not sure though, that you'll find enough volunteers/experts to make it work.

7/12/2010 at 6:11 PM

I think the script idea is a good interim measure until geni internally develops a solution. It wouldn't need to be a perfect automated process, just something to make it faster and easier for those interested in preserving data to correct errors. So why not just start with a simple greasemonkey script that allows you to take a snapshot of a profile, and then when you revisit a profile, highlights changes and allows you to reset it back with one click? It won't fix additional spouses and other complex problems, and it would provide notifications of changes and probably won't work from the tree view since that's flash, but it would at least make it faster to make some changes...

you're forgetting one thing - how ornery people can be. If just anyone can QC tag their profiles, then people will do so, just to "protect" the data they copied off the net. And yes, I commiserate regarding the amount of shit-data out their. Having any OTHER person won't work either (plenty have multiple accounts).

As to routing or grouping people, I already tried that at one point. If you go to Geni's Wiki - which I see you've just joined, you will see on the very first page a link to Merge Collaboration - inviting people to create "merge projects", and by stating their areas on interest, find people to work with. That page has been viewed a grand total of... 2,309 times, in 16 months. So while a LOT of projects were created, they are mostly solo efforts. The Wiki in general has petered out, with about 10 active users (*including* Geni Staff).

In reality, if you choose to become active in merging [part of] the historical tree, you will find yourself in contact with a relatively small group of power-mergers. We used to have an informal "merge group", with dedicated Discussions. Due to recent improvements Geni made (some per request), like merging through a shared Collaborator and then auto-accept, the need for the group was much reduced, AND it has been superseded by the open-to-all Collaboration Pool - If you like, I can suggest a "short-list".

Steven, I now have a much better idea of what you meant by "script". I like it!
Yes, that probably wouldn't be to hard to do. We could even do it independently of Geni, if we stored the snapshots on our own computers, or on some shared server (wee bit more work). Power-user Erin Spiceland is very handy with such things. She had created a zombie killing list, before Noah came along with his official one ( ).

The only possible problem I see is if 2+ people have different ideas of what the profile should look like, and maintain separate snapshots. You'd have a constant revision-war. :-P This is true at present too (and happens), but it takes much more effort to keep-up. For this reason, I would advice against making this a freely available script. But then I also hate "in-groups" even when I'm "in". I have NO idea how Geni's "soon" upcoming feature for profile revisions will avoid this issue.

In General, regarding management, you CAN request to be added as a manager to any profile. You can even ask the current primary manager to make you primary. Many people are very accepting of such requests, especially if they know you, and trust your care-taking of this part of the shared tree.

here's my "short list". These are OF those Collaborators of mine that I find most useful to me, either because of the sheer number of Collaborators THEY have (some have thousands), the number of profiles they manage (probably have the top five on Geni in this list), or for being so very helpful and knowledgeable about various parts of the tree. I probably missed a few, but also wanted to keep it short! People really should just join the Collaboration Pool and be done with it.

Terry Jackson (Switzer)
David Lee Kaleita
Sean Patrick Feeney
Private User
Gwyneth McNeil
Private User
Günther Kipp
Cecilie Nygård
Private User
Private User
Gene Daniell
Private User
Pam Wilson
Kris Hewitt
Private User
Lúcia Pilla
Erin Spiceland
Henn Sarv
Martin Severin Eriksen

Private User
7/17/2010 at 10:07 PM

Everyone has their own "area" they know best.

For my part... I'm best in the Alden/Mullins tree, as that's where I spend most of my time.

I would be completely lost in the Plantagenets though.

Perhaps the best way to find the answer... is simply to ask.

Jason P Herbert

Showing all 8 posts

Create a free account or login to participate in this discussion