Re: [tcpm] urgent data draft (draft-gont-tcpm-urgent-data-01.txt)

[Joe Touch <[e-mail]>]

-----BEGIN PGP SIGNED MESSAGE-----
Hash: SHA1

I'm wondering about #2. It's worth noting that implementations don't
follow the specs, but I'm getting increasingly concerned about
continuing to publish docs that say "implementations don't follow spec"
without actually either saying "and the spec is hereby changed" or "and
implementations are noncompliant and should be fixed".

While I agree that both #3 and #4 are the most important
recommendations, I propose that the WG should decide how to address whether:

	1) decide whether to update 1122 and clarify 793 (as
	standards track)
	or to reiterate 1122 and indicate existing implementations
	noncompliant

	I don't really care which of these is picked; given
	#3 and #4, it wouldn't really matter much, but I do
	feel strongly that the WG should NOT ignore this aspect.

	2) decide whether the doc is standards track (if changing
	1122) or BCP

	IMO, BCP is required based on #3 and #4 (i.e., it's not
	about whether existing impls. are incorrect, so much
	as advice on how to proceed or the future). If 1122
	is changed, then the doc needs to be standards track,
	which covers the "BCP" advice in 3 and 4 anyway.

Joe


Gorry Fairhurst wrote:
> I supported this in the meeting, and still do.
> 
> I think a statement that this not recommended for new applications (3)
> is good, combined with (4).
> 
> Best wishes,
> 
> Gorry
> 
> David Borman wrote:
>> I think I forgot to follow up on this, sorry!
>>
>> So, with my WG co-chair hat on:
>>
>> The agreement at the San Francisco IETF meeting on the Urgent Pointer
>> definition was:
>>
>> 1) Adopt this document as a WG item
>>
>> 2) Change the definition of the Urgent Pointer (defined in RFC 1122)
>> to match the definition on page 17 of RFC 793, which is what most
>> implementations use.
>>
>> 3) New applications should be discouraged from using the Urgent Pointer
>>
>> 4) TCP implementations still need to implement the Urgent Pointer for
>> existing applications that use it
>>
>> 5) All applications that do make use of the Urgent Pointer must use
>> the SO_OOBINLINE socket option to keep all of the data in sequence;
>> applications that don't use SO_OOBINLINE and continue to use the old,
>> broken BSD implementation that actually removes bytes of data from
>> data stream are out of scope for the IETF, since that is not part of
>> the TCP protocol.
>>
>> Please respond with whether you do or do not support adopting this as
>> a WG item, even if you were at the meeting, so that we have a record
>> on the mailing list.
>>
>>
>> Now with my WG chair hat off:
>>
>> I support adopting this as a WG item.
>>
>>             -David Borman, TCPM WG co-chair
>>
>>
>>
>> On May 19, 2009, at 10:01 AM, Fernando Gont wrote:
>>
>>> Hello, folks,
>>>
>>> I'm planning to work on a revision of the urgent data I-D anytime soon.
>>> I was wondering if there are any plans for proceeding with this I-D.
>>>
>>> There was some discussion on-list about TCP urgent data, and IIRC Dave
>>> Borman had suggested (hat off) that this I-D be adopted as a WG item.
>>>
>>> Thoughts? Plans?
>>>
>>> Thanks!
>>>
>>> Kind regards,
>>> -- 
>>> Fernando Gont
>>> e-mail: [e-mail] || [e-mail]
>>> PGP Fingerprint: 7809 84F5 322E 45C7 F1C9 3945 96EE A9EF D076 FFF1
>>
>> _______________________________________________
>> tcpm mailing list
>> [e-mail]
>> https://www.ietf.org/mailman/listinfo/tcpm
>>
>>
> 
> _______________________________________________
> tcpm mailing list
> [e-mail]
> https://www.ietf.org/mailman/listinfo/tcpm
-----BEGIN PGP SIGNATURE-----
Version: GnuPG v1.4.9 (MingW32)
Comment: Using GnuPG with Mozilla - http://enigmail.mozdev.org

iEYEARECAAYFAko/wcIACgkQE5f5cImnZruB6QCeNm660Aa7LTdeFknidCvslG18
xwwAnjnoDMRcW353Oi4PuVznssHshS+i
=OHXZ
-----END PGP SIGNATURE-----