The IDs are there because I needed to check if the font I chose later showed all the characters including the diacritics. If someone wants to change the font in his/her epub now, they can still easily look up all the instances where they appear.
As there are no complaints yet, I assume the IDs are not a problem with current readers. Considering future models sounds too hypothetical to me. And I have read several times about Sigil complaining about epubs without anyone having a problem opening them on their readers, so right now I don't intend to update it.