While the rfc2307 schema incorporates attributes, for naming users and groups, for which equality matching is done in a case-insensitive matter, client systems usually have the assumption of case-sensitivity built-in.  When a user manages to log in as "Jimbo", but a group references that user as "jimbo", weird things will happen (this includes cache misses, authorization problems, and perhaps some crashes, too).

I'd like for us to add some rules to normalize incoming data so that we can at least shield administrators from unintentionally creating a situation where case-mismatched data is in the directory, because that's much harder to detect and compensate for at the client.

Whether we do this in the administrative UI or at the directory level with a plugin (the latter might become a hurdle for mass-import in migration cases, so I'm not too sure about it) is still up in the air.