I'm wondering what percentage of newly created profiles for people born before 1850, say, will end up getting merged into existing profiles. I keep stumbling over profiles that need merging, and too often I can't see any reason why someone can't come along later and create the same duplicates all over again.
The eccentric behavior of the WikiTree search engine could be part of the problem. In suggesting matches for a profile I propose to create, sometimes the search suggests people with similar names but born impossibly early. Sometimes it suggests a living person, born impossibly late. But evidently it can still miss actual likely matches. For instance...
Before their recent merger, when I searched for Nicholas Brink, Nicholas Brink-xx showed up and Nicolas Brink-yy did not. When I searched for Nicolas Brink, Nicolas Brink-yy showed up and Nicholas Brink-xx did not. No wonder the second profile manager didn't get warned off.
Also if I searched for Leonard (Cool) Brink, Leendert Cole Brink-zz didn't turn up, and if I searched for Leendert (Cole) Brink, Leonard Cool Brink-aa didn't turn up. A merge is pending.
The "Browse Matches Tool" might be helpful, but it only works for searching for matches for profiles you've already created, not for ones you're planning to create.
I'm not sure what software tricks would be needed to get the search algorithm to include more variant spellings. I guess it does, to some extent, but not enough.
Possibly easier to arrange would be to keep track of names that were merged and be able to list them. Thus if I search for Lambert Brinck, wanting to create the one that was baptized in 1635, first I'd get the usual list, but there'd be a reminder to click the button to see the list of all the profiles that were merged into the profiles on the first list. These names would include their numbers and be linked to the profiles they merged into. Thus I'd see my Lambert Brinck on the second list and discover that he merged into Lambert Huybertsen-4 on the first list, whereas "Lambert Brinck" born in that year wasn't on the first list.
Or if I wanted the Lambert Brink born by 1702 (he married Rachel van Garden), I'd still be out of luck because at present the search engine doesn't find him unless you ask for Lambertus Brink. Therefore it would be good if the 2nd list not only included the profiles that disappeared into the first list, but also all the profiles using the same names as on the second list but that disappeared into profiles not on the first list. Probably several profiles were merged into Lambartus, so I'd likely find the Lambert Brinck I wanted there.
Letting people see how many profiles have merged into some of these popular profiles might inspire people to try harder to figure out how to prevent duplication. Sometimes duplicates contain complementary pieces of information, but more often they waste the later profiler's time, searching for information that's already been found.
This list of merged profiles would be even more helpful with the Depuy family where there's not just two main spelling variants, but maybe a dozen, and the search engine won't let you see them all at the same time.
Being able to ask the search engine to provide names born within a certain range of dates, say a decade or two, would also be helpful.
Again, it would be interesting to know the answer to my first question, about what percentage of profiles end up getting merged. Seeing how this percentage changes over time, say comparing this six months' to the last six months', would give us a measure of whether the remedies that are being tried are actually working. ...Although, since mergibles often go undetected for long periods, these percentages might not be nearly accurate until they're at least 3 years old, say.
