10/1/00
None
10/2/00
Thread: Provisional responses in other direction
Q. UAS may send ISUP bodies in 1xx and 200 responses. What if the UAC
wants to send additional ISUP before final response is received?
A. Gonzalo Camarillo: use INFO. Cliff.Harris@nokia.com:
this has been
discussed before: caller may not have enough data to send INFO (no tag,
or Route). See
http://lists.bell-labs.com/pipermail/sip/2000q2/000006.html
for previous
discussion.
Thread: Second branch parameter
Q.Section 6.46, August bis talks about branch ID consisting of two parts
but doesn't say what the delimiter is. Also, two parts a SHOULD, not a
MUST
A. "second branch parameter" is by definition part of the whole branch
parameter,
thus you just use the whole branch parameter to distinguish. Just update
the wording to avoid confusion.
10/3/00
Thread: Error in Via received BNF?
Q. The Sept 4th bis, Section 6.46.2, says that, "...If and only if the
source address and port differ from the sent-by address, the proxy also
includes the source port in the received parameter." But the BNF for
received tag has no port.
A. The description is incorrect, only the address is checked. The text
is updated.
Thread: SIP ISUP MIME draft
Q. Any comments on the draft-ietf-sip-isup-mime-04.txt?
A few questions came up:
Q. Can Content-Encoding be used with ISUP payload?
A. No, it's in binary
Q. Is Content-Disposition a mandatory field?
It's optional but there should be default. No default is currently
specified. "session" is not appropriate. New Content-Disposition of
"signal" is proposed and accepted.
Thread: centralized vs. decentralized architectures
The name notwithstanding, the question was about assigning ad-hoc
addresses to downloadable SIP phones. Joshua_Fox wanted to solve the
problem by anonimizing the caller. JR response that it doesn't make much
since "network elements that provide service to this payphone would
authenticate it"
10/4/00
Thread: Registering non-SIP URIs
section 6.14 has one paragraph about 'REGISTER requests': "Other
[non-SIP] URI schemes have no expiration times." Would be nice to move
it to 4.2.6, where REGISTER request is described. The changed was made
to a later version of bis.
Q. Why is the default expiration interval 1hr? A. No reason, it's fairly
arbitrary
Thread: ipv4 addresses in sip bnf
Q. Why IPv4address BNF in SIP is different from the one in RFC2327
(SDP)?
A. Because it's inherited from HTTP.
10/5/00
None
10/6/00
Thread: New draft on SIP registration
Comments on the new submitted draft:
Q. The draft talks about using Authentication-Info but 2543bis says that
this header is not used in SIP.
A. We may want to allow Authentication-Info; besides, we need a
mechanism to protect at least some parts of the header via Digest.
10/7/00
None