Mark Crispin wrote:
Yes, it is odd, but it seems to be more useful.Can we update the next revision of IMAP4rev1 to recommend this behavior?
A year ago I've done the same change as Tim suggested (a.b.c.d) in M-Store, but I agree that Mark's interpretation has its advantages.
I'm sure that we should recommend a.b c.d NOT evaluate to a.bc.d, which strikes me as never the right thing. I hope we can at least remove that recommendation, but leave the "implementation-defined" language in place, and perhaps the a.b.c.d example.
Tim
