>> Ah, so the problem arises when un-quoted and quoted usage is both used over
>> time. Why would anyone do that or why would a tool do that?
>Why is a hard question to answer.

Well, not all too hard, I can easily think of two situations:

1) Over time, reserved words are added to the database - e.g. I remember that we used TYPE as a fieldname (in retrospect, it was a silly choice of us, but having all fieldnames in Norwegian, you don't think about the meaning of the name in English - even if it is identical. It was only former problems that made me insist on not getting SET as a username ;o). Doublequoting such field name may make some programs continue to work unchanged, without forcing you to doublequote (or use correct case in) all fields in other programs.

2) Some tools may want to simplify the creation of tables as much as possible for the user, whilst providing maximum functionality. I'd say that would normally mean using unquoted identifiers (so that person, Person and PERSON all referred to the same field or table), but still add quotes when the user wanted fieldnames that weren't allowable as regular identifiers - e.g. "Person id".

Set