Re: Comments on draft-ietf-conneg-feature-syntax-01.txt

[Graham Klyne <GK@xxxxxxxxxxxxxx>]

At 08:23 13/11/98 -0500, Al Gilman wrote:

>I also believe it might be useful to state explicitly that q=0
>should be treated as "categorically unacceptable."

I know we've both expressed an opinion on this.  I shall try to be more
precise about my view in the hope of eliciting a consensus:

(a) I don't believe that adding such a constraint would add any value to
the -syntax- draft as it stands.

(b) I don't believe that the absence of such a constraint in any way limits
what can be achieved by implementations that may wish to apply this view.

(c) to make this statement would 'fix' a relationship between the q-value
and the predicate which I think could create some additional unnecessary
complexity.

(d) making this statement would impose additional constraints on the form
of semantics that could be applied to q-values.  Specifically, the expression:

    (& (foo);q=0 (bar);q=n )

would be constrained to have a q-value of zero.  This may be reasonable,
but I don't see that it's always necessary.

In making statements (c) and (d), I am assuming that your statement means
that any expression of the form:

    (foo);q=0

must be interpreted as equivalent to (FALSE).

#g

------------
Graham Klyne
(GK@xxxxxxx)