Re: [sipcore] WGLC: draft-ietf-sipcore-name-addr-guidance

[[e-mail] (Dale R. Worley)]

Yehoshua Gev <[e-mail]> writes:
> The examples of:
>    Refer-To: sip:123@host?Replaces=1111
>    Refer-To: sip:123@host;user=phone?Replaces=1111
> will all be syntactically invalid using the interpretation proposed by the
> draft.
> I think this is ok and corresponds to my current understanding of RFC 3261.
>
> However, in the above thread, it seemed to me that on first sight, both Dale
> and Robert thought that those examples are valid.
> After the discussion there, I think (IIUC) they agreed that those are not
> valid.

You have to be careful about the terminology.  Within the context of
draft-ietf-sipcore-name-addr-guidance, there is the BNF, and then there
is an extra-BNF constraint.  So what does the term "syntactically valid"
mean?  In the context of parsing as it is used in computer programming
languages, "syntactic" is usually used only in regard to what is
expressed in BNF.  So from that point of view, you could say that the
two examples are "syntactically valid" but that they are invalid
relative to the non-BNF restriction.

On the other hand, you could say that "syntactically valid" means that
it passes every validity check that can be implemented without reference
to external information sources.  In that case, the two examples are
"syntactically invalid", because they violate the non-BNF restriction,
and determining that doesn't require knowing anything about the domain
"host".

That ambiguity is why I used the term "parsable *by the BNF alone*", as
that is not bedeviled by the exact definition of "syntactically".

Of course, the questions you raise about what examples should be
discussed in the draft are still valid.

Dale