Through the painful story of our family tree, not yet completely solved, I will try to explain that, in my opinion, there are 2 methods for merging big trees: a crazy one and a good one. And I would like to have the opinion of some users who experienced similar stories, and/or who acquired experience in merging big trees in different situations.
Our family started with 3 independent trees (A, B, C) for the descendents of a historical ancestor. I have been nominated by the group of descendents to manage a pro account and the C tree. C counted about 750 profiles, with by far the highest quality of data, because this data had been introduced by each of us individually. B counted about 500 profiles, with lower data quality, because the data was introduced only by 2 of our cousins. Bad luck for us, these trees were of high interest to the curator, who was exactly working on some of the involved ancestors. Her tree D was already connected to the big tree, and also comprised a big number of duplicates with C, with lower data quality, at least for the most recent generations.
Using different stratagems (I will not describe here to make short, and also because each of these stratagems can make a different post here for the future) the curator penetrated our family group, in the 3 trees, and performed several wild tree merges without our consent, and without any preparation. This is against the recommendation that Geni itself writes in its help, and also in the small red announcement that is displayed below each profile merge, warning that profile merge could lead to a tree merge and numerous duplicates.
The number of duplicates coming from A and D is not known, but it happens that we perfectly know about B: B duplicated about 400 of C profiles, with lower quality data. My strategy as manager was that each of the cousins suppressed the duplicated profiles in A and B before merging with C, and then, study carefully the possible duplicates between C and D, before doing this last merge. But, by her brutal and blind action, she did not leave us the time to do it.
As a result, we got a complete chaos in our tree, hundreds of duplicates, merges to complete, tree or profile conflicts to solve! Number of such incidents to solve reached a peak of 362 on march 5 and is down only to 109 March 30, despite 534 repair actions that I performed on the tree (merge, suppressions, conflict solving, etc….).
The good method:
Suppressing a profile with lower data quality takes about 3 s in the tree mode. For the tree with 400 duplicates with lower data quality, it would have taken maximum 20 min to do all the job! And this job could be shared, each cousin doing it in his own tree.
The crazy method:
Solving duplicates take much longer: 10 s to display profiles, study them, decide what to do, 3 s to merge, then again 30 s to 1 min more to solve the resulting tree conflicts and/or data conflicts. In practical terms, it took me more than 30 h over about 18 days and it is not yet finished. And this job cannot be shared, as my pro account at this time was the only one.
So, for me it becomes absolutely obvious that, when you know in advance that a big % of the candidate tree to merge has lower quality data than yours, either you should reject the merge, or, if you accept it, you should negotiate first of all elimination of all duplicate profiles bearing lower data quality.
For this reason, we had a hot discussion with this curator about this, and a lot of other issues, each of them I will come to in my next posts. All this leading to nothing less than 6 successive claims to Geni, still in process.
If as a user you experienced similar stories, and/or you acquired experience in merging big trees in different situations, I would be interested in having your opinion.
