> It because the spec is crud.

It is because it has been well thought out to accommodate case sensitivity
as well as a good measure of backward compatibility without allowing the
least bit of ambiguity.

> if using dialect 1 then any case is converted to uppercase and therefore
> accepted.

This is legacy behavior. Not much to do about this in version 6.

> if using dialect 3 then the case must match exactly and identifiers must
> have quotes.

Thankfully it is so in order to avoid ambiguity.

> However, if the identifier is all uppercase then it cant have quote....

note

> CANT.

No, it can. Unfortunately this is where a little ambiguity can come in if
the component vendor isn't real careful.

> To top it all, every component writer makes different assumptions about it
> converting case and adding/removing quotes in a non-structured fashion.

The rules are very clear. Some chose to follow them and some chose to abuse
them.

Regards,
Jason Wharton
CPS - Mesa AZ
http://www.ibobjects.com