The LNAB was the main reason why I was a long time reluctant to join WikiTree. I eventually got over it, and am happy I did, because I've learnt a lot here. WikiTree has other issues, but I can live with all of them. They are the issues of all online collaborative endeavours, which did not compromise the growth and success of Wikipedia, for example. But LNAB keeps getting in my way, and on my nerves. And it will be the one and only reason why I quit, if I do.
And it could unfortunately be the major stumbling block on WikiTree's path towards achieving its universal mission.
I could go for hours listing the dire consequences of this basic feature which is in fact a serious congenital bug. And I'm on the verge of going away because of it. The problem being, where to? WikiTree is technically flawed, but there is nowhere to go with a similar spirit and community. So let me sum up once for all why LNAB is a bug, even if it's useless because I know the answer : c'est comme ça et on n'y peut rien!
1. Using a name as a basis for identification in a non-starter, and the more so in modern information systems. Even librarians, who have a looong tradition of maintaining "authorities" with "preferred names" and "alternate names", and the like, who have munched ad nauseam over the issues linked to naming and indexing for centuries, have eventually understood what LaoZi wrote in six characters very long ago, 名可名，非常名. Librarians came out with shared identifiers (dumb numbers) around which they gathered all the variety of names for a person, or for that matter a place, a book, or any subject a library catalogue wishes to reference. Each great national library did it, and then they sit together to create VIAF, a ring to rule them all. I'll be back with an example below.
2. Why "last name" and "at birth"? Looking at some recent and less recent questions in this forum, the mandatory LNAB to create a profile leaves to ridiculous debates and waste of time and energy. And what if there is no birth record available? And what if the birth record had it wrong, and was corrected afterwards? What if the siblings in the birth records have different family name? And what if the person belongs to a country, a language, a culture, a period, where the very notion of "last name" does not make sense? Those things happen all the time, and I read in a recent debate here that people just do not create profiles because they can't figure the LNAB.
3. Adding to the above, if you make a mistake when creating the profile, even a simple typo, of course you can change it, but it means creating a new profile and merging, and we are said not to do that too much, because it creates a lot of redirects in the data base, which is an overload on the system. A problem perfectly technically understandable, but from a end-user perspective, is seen as a big bug.
4. From a search point of view, LNAB is often not the most frequent name under which the person is known, and people searching in WikiTree, directly through the internal search engine, or from an external one, are likely either to miss it, or have hard time recognizing that they have indeed found the person they were looking for. I had this experience a lot of time searching famous people in France history, for whom I had to try several names before finding them out, or go through some member of their family I know was there etc.
When you put all that in the context, say, of French aristocracy, where the syntax and semantics of names are both very codified and of course suffering many exceptions to its own codes, it's bound to be a total failure. The adaptation rules chosen to make those convoluted names enter the data base structure are just pathetic, and will send any serious french genealogist de bonne famille to roll on the floor laughing, or cry in despair, but by no means will he take the whole thing as serious, and never ever join us. If any true French aristo with seize quartiers de noblesse is around here, could (s)he please step forward?
And moreover most of those profiles have been created by people with only vague notions of those naming subtleties, but who happen to be far descendants of an aristo, or think they are, they have it wrong. Which should not be a problem, because the Honor Code says We know mistakes are inevitable We don't want to be afraid to make them. But if mistakes relate to LNAB, you are not welcome to correct them because of point #3 above.
I stumbled on one of those, in reference, and I put a comment on the profile that Henri de La Tour d'Auvergne means, as far as I understand it, a given name of "Henri" to a member of the "La Tour" family, of whom fief was "Auvergne" at some point (a province of France). It was before "La Tour d'Olliergues", Olliergues being a city in Auvergne. You see there that the family has gained power. More at https://fr.wikipedia.org/wiki/Maison_de_La_Tour_d%27Auvergne
So, all people in this family, if we were to apply at best the silly LNAB rule, should certainly have as LNAB "La Tour", aka "d'Auvergne" for our Henri, best known in Histoire de France as "Turenne".
The answer by Isabelle, after we had a private exchange about it, reads : yes it's a mess, but what could we do, because of all above listed reasons? My answer is : get rid of LNAB, rethink the whole naming issue on a really universal basis as librarians have done at some point, to allow any local naming convention, save the data, and migrate to a new data base. WikiTree will have to do that at some point, or die out, or stay forever a permanent frustration for all those here who are dreaming its dream. If change has to be done, the sooner the better. Migrating 20 million+ entries is not easy. When we are at 50 or 100 million, it will be worse.
Going back to VIAF : https://viaf.org/viaf/59190793/ is our man, see how different libraries have chosen different "preferred" ways to name him ... and the 218 (to date) "other forms of the name", just missing the one chosen by WikiTree.
[edited to get rid of rude language, thanks Julie for pointing at it]