I totally understand why it wouldn't be practical to do something about the issue but in many cases, diacritical marks are required to differentiate between words. For example, in Portuguese Grandpa is "Vovô" and Grandma is "Vovó". Another example is "para" and "Pará", "para" meaning for and "Pará" being a state in Brazil. Another great example is "coco" and "cocô". "Coco" means coconut and "cocô" means something that you will just have to find out for yourselves if you really want to.
If making the system recognize accented letters as being the same as their unaccented versions is not possible then I don't think using diacritics would be a good idea, just as zuzu said.
Also, if there is ever the need for someone to go and fix names in the database, I'll be glad to do it. I have and still do fix names on Google Maps all the time. I think it's fun actually.
I think Vietnamese is probably my favorite language to correct.