[Data-modeling] how much work done on modeling of personal names -- even for surname + given name?

[glenn mcdonald]

>> me: The big advantage of [just a sortname] is that it keeps you  
>> out of
>> the quagmire of modeling all the internal semantic complexity of
>> worldwide naming patterns, but still allows you to model the sorting,
>> which is the thing you most often care about.
>
> Tom: The disadvantage is that it enforces a single global collating
> sequence.

Well, it doesn't "enforce" anything. It provides one more sorting  
option than you have now, so you at least have straight lexical sort  
(the current name) and the most culturally appropriate name sort (the  
new sortname). And by definition you can populate this new field with  
*something* for all people, and then fix the ones that need fixing.  
Seems to me like having the culturally-appropriate sortname is a good  
thing even if you also have the name split up into parts...