You could lead the way and make your historical records and smart matches available by a paid API to other companies, and those companies could pass on the fees to their users that opt in to that service, say some combination of per-company, per-user, per-search, and per-result.
There is no need to bottle it all up, this is how things work for many big companies out there (in the wide world outside of genealogy)
As you may appreciate, MyHeritage and Ancestry are competitiors and thus there are certain collections of data that are unavailable to either company due to the collection being held by one or the other.
If you however have a copy of a record from an Ancestry search that is a JPG image saved on your computer, you can then quite easily upload this document image to the specific individuals profile within your MyHeritage family tree.
I have a problem where I get smart matches to someone with the same name, but all the other family members are private, so I have no way to confirm if it is really a match.
Considering I've had a number of wrong matches in the past I've had to reject, I have to be careful about this.
Say I match my "Bob Cobb" to someone else's "Bob Cobb", and I have his wife as "Sheryl Cobb" but the wife in the other tree is "<private> Cobb" - can I assume that "<private> Cobb" is not Sheryl? Or it could just be that they have "Cheryl"or "Sherry" or "SHERYL DIED IN 1974" (sigh!).
So far nobody I've written to on this site with a match like that has responded :/
I've had this problem a few times. Someone will be called "John Andrew Smith" and his brother will be "Andrew John Smith" and smart constantly brings up the matches swapped around. It means I can get 100 smart matches to one tree but I can't simply click "accept all". And the strange thing is that the same wrong matches will come up several times in one tree. Ugh!
The birth years of all the entries are correct, why doesn't that help it resolve the right person?
Does rejecting the matches have any effect on your algorithm? Does it learn from it's mistakes?
I get a lot of smart matches and instant discoveries for branches of the tree that are not very important for me to build out. I think you guys could save yourself some processing power if you allowed me to indicate that those branches should only be revised by the smart matcher / instant discovery technology on a very occasional basis.
I guess you could automatically determine which branches I'm most interested in - my ancestors and their descendants. I'm less interested in the distant cousins of someone married to my distant cousins. I'm less interested in the ancestors and cousins of someone's ex-wife.
(However, that could easily not be the case for everyone, just because a relation is distant on the tree doesn't mean that they are distant in person, so those tree maintainers may wish to change their preference)
It's fine that those matches come up, but I'd prefer not to deal with those discoveries/matches very often if possible.
Perhaps that is already the case and you have the CPU cycles to spare because you couldn't find anything for closer relations, I don't know. It just seems like I've done an awful lot of work reviewing matches for ex-wives' distant relatives.
When you confirm a match for an individual you often get a screen giving you the option to import "more complete'' details. Great.
However I often use the Confirm All feature, and so I don't get this. Is there any kind of subsequent automated checking for missing details in my tree that are present in matched trees? Other than additional persons provided by instant discoveries, I mean.
When a person is matched in various trees by several users, but there is a conflicting detail, e.g. some people say he was born Jan 1827, some people say Mar 1827.... what becomes of this?
I guess nothing.
I reckon something should happen though. Users should be alerted to conflicting details, and perhaps a comment thread should appear about those conflicts that the maintainers of the trees can participate in to resolve the issue.
Rather than deleting matches, you should confirm or reject them.
Rejecting gives you the opportunity to review matches later as your research progresses, even if you decided they weren’t good the first time around.
It sometimes happens you’ll receive a match that upon first look seems completely wrong yet turns out to be correct at a later date. It would be a shame to lose that connecting thread if the match had been deleted in the first case.
Rejecting, rather than deleting is meant to help you retain research leads that can advance your work when you get stuck.
If you would like to reject Smart Matches, please read one of the following FAQ article: