[Date Prev][Date Next][Thread Prev][Thread Next][Date Index][Thread Index]
Re: Working group last call for -reg-
What do you think of what Ted did with this in the "wordsmithing"
he described just now? It looks good to me.
to follow up on what Koen Holtman said:
> From koen@xxxxxxxxxx Fri Jul 24 11:40:46 1998
> From: koen@xxxxxxxxxx (Koen Holtman)
> Message-Id: <199807241540.RAA09765@xxxxxxxxxxxxxxxxxxx>
> Subject: Re: Working group last call for -reg-
> To: asgilman@xxxxxxxxxxxxxxxx (Al Gilman)
> Date: Fri, 24 Jul 1998 17:40:34 +0200 (MET DST)
> Cc: ietf-medfree@xxxxxxx
> In-Reply-To: <199807191347.JAA18387@xxxxxxxxxxxxxxxxx> from "Al Gilman" at Jul 19, 98 09:47:16 am
> X-Mailer: ELM [version 2.4 PL23]
> [Not a discussion of the -reg- draft anymore...]
> Al Gilman:
> >However, it is not safe to treat "absent" and "false" as the same
> >thing. There is a logical "does not apply" case as well as
> >unknown. I believe it is impossible to guarantee that "False" is
> >safe in general, and so the rule should be:
> > Any agent in any protocol which does not recognize a
> > feature tag shall ignore the property typed by that
> > tag. That is to say it shall process as though the
> > property were absent.
> I would interpret this rule as specifying a default too, though one at
> a higher level of processing. And like any default, it has cases
> where it can be dangerous: if a recipient which does not understand
> the explodes.unless.preprocessed.with.the.xyzzy.filter tag processes
> content labeled with this tag as if the tag were absent, bad things
> will happen.