Lack of tree import feature is pretty bad.

I can actually see where I could attach my end branch here but the top part I'm pretty sure I see a number of mistakes so I don't this system necessarily helps anything. If anything it might make it worse, since less people are likely to be willing to bother trying to do work here too.

Upload your gedcom to myheritage and download the smart copy app to copy it over to geni

That appears to require pro subscriptions to both services to use. Some make it sound like you have also walk it through leaf by leaf too.

Not sure what the deal is with the ban on GEDCOM uploads here. It 100% does NOT lead to more accurate info ending up on here. Seems like some bizarre marketing scheme to me t try to force people to use this is the base tree site on which to do their primary tree building on.

I guess I can test out the free trial here and see if the 23andme special unlimited leaves at MyHeritage is enough to get the tool working over there and see whether the chrome smart copy tools is workable or not.

The reason for the imports of gedcoms is not allowed anymore is that it made to many profiles of the same persons, specially in the medieval period. And that again produced a lot of work merging all of these duplicates together.

If you or anyone else are planning to use the method Jason talks about, please make sure to merge in all the duplicate profiles you add, or make sure that there are no duplicate of the profiles you are adding to Geni via Smart Copy before you copy them over to Geni.

I still think it's extraordinarily counter productive to not allow GEDCOM uploads. It is simply not feasible for me to hand-transfer over my 5400+ person tree. Instead I have a ragged bit of work there which does nobody any good at all.

And far from this no GEDCOM uploads making things more accurate here, I'd dare say it makes for far more errors. People may fix mistakes elsewhere but then find it to much a nightmare to re-do and fix things here by hand you get all sorts of old out of date messed up trees here.

And in the area most of my family is from I see some trees with far more huge errors than I do at say MyHeritage where GEDCOMs can be uploaded so....

And don't forget that many people come from areas where you won't even be tracing into the medieval times at all or be lucky to find more than a few matches to other trees out of every few thousand entries....

So my tree here still remains a tiny runt, perhaps even slightly muddled, while on say MyHeritage it has tons of good research being presented.

And I see a few large trees here that I know for a fact are at least 50% utterly wrong....

Only reason you should need to add 5000+ profiles to Geni now is that your genealogy comes from a place where there are few profiles added before.

Most of the time you will hit profiles of the same person within a few generations of the ones living today, and then you should work on those profiles instead of adding your own.

It reminds me: The last time I was in a discussion with a person claiming that GEDCOM import really was needed because he had thousands in his personal tree the practical conclusion was that he only needed to add his parents.

If you make your tree on MyHeritage freely available, and ask for a volunteer to help you copy over from your tree to Geni with SmartCopy, I'm sure something will happen.

Out of curiosity: Which large tree have you seen here that was 50% utterly wrong?
As far as I know, there's no significant trees beyond the Big Tree left, and it's unlikely that its 100+ million profiles has 50+ million obvious errors - if you meant to say "I have seen areas of the Big Tree with very many errors", I'd most certainly like to know where they are.

Harald, you forgot to tell those that use SmartCopy that they still need to merge in all the duplicates they make. That is an important information, or else you, me and other curators usually have to do the merging since the person that used SmartCopy doesn't fancy doing the mergeing of several hundred or thousand of profiles.

That is also why I really don't like the advertising of SmartCopy since it leaves us with the same job as adding a gedcom would do.

I think that it is theoretically possible to make an import feature from myHeritage tree which will not create a duplicate tree. There are already Smart matching and Record matching features functioning over myHeritage/Geni trees which are quite good and usually founds that correct matches. So I think it is possible to code the import feature such way that if there is a confirmed smart/record match, no additional profile is created on Geni, instead the data from myHeritage is put as alternative dataset to the existing profile (like it is working now when two Geni profiles which differs in data are merged together). This way no duplicate tree is created.

It will help enormously in cases when someone have a tree from the region which is not presented much at Geni. For example my own myHeritage tree have slightly less than 6k profiles. Calculated Record matches shows that I do have about 350 matches with Geni. 5,5k records to copy even with a help of a tool is quite a work. Especially that it most probably cannot copy the attached photos (I had plenty of scans of the BMD records).

Remi Trygve Pedersen that was mentioned previously. I make the assumption that anyone who volunteers to do importing on Lauris' tree will alreday know that he has to take care of the merges. I may be an optimist :-)

Тигран Юрьевич Лачинян this might work, if the import process halts when it finds a match, and does not resume until a human has rejected the match. In most cases of a match that smart matching is able to find, there are 10 more matches that smart matching didn't find in the surrounding area, and this would have to be merged after approving the match. What would happen to the rest of the import isn't clear.
Not a *small* matter of engineering.

Note: SmartCopy is able to copy photos, but this is turned off by default because you need to be sure you have the copyright permissions needed for the photo you are copying.

Yes, the process I thought about must be interactive (and probably pausable and resumable), but still it would be much more automatic than the currently available options.

Every time a matched profiles are merged the process may propose to merge the surrounding area using the same mechanism which is already working when merging two profiles of Geni tree - you may drag and drop the appropriate profiles one over another to link them. Of course if some of surrounding profiles are already matched, they would start in a linked state.

I think that the process at start may create the hidden profiles, visible only to the user himself and possibly for his collaborators (not appearing for example in the Geni search for other users). After the user (or collaborators) will go through all the matches and merge or deny the match, all remaining profiles from this import will be made unhidden. If the user will not end the process of merging or denying matches, after some time of inactivity the hidden profiles will be deleted. Some problems may occur if two or more users will try to make an import regarding the same part of the big tree. I think that the best resolution would be to encourage them to collaborate or in case they will decide against it, to block the possibility of ending their import processes simultaneously - so if one of them will end his import, others will need to wait for tree matches being recalculated with their hidden profiles and then they must merge or deny the matches with the new profiles to avoid the situation when the simultaneous import creates several duplicates of a profile which previously was not in the Geni.

Additionally I think it is possible to create some sort of links between myHeritage profiles and Geni profiles. Such links may be created either by confirming the match, or automatically during the import. Next time the process may smartly skip the already merged profiles unless something changed in data or they have new connections (children, parents, spouses, siblings).

I agree that it is not a easy task, but if the process will be designed in such way that it will use as much as it possible the already available features, it will save some work.