Re: Adding parallelism? (was Re: Cost vs. Benefit of Real-Time Applications and Infrastucture Area)

[Date Prev][Date Next][Thread Prev][Thread Next][Date Index][Thread Index]

 



Pekka Nikander wrote:
So, I am hearing here two conflicting messages:

1. We can't consider a "horizontal" split, since we need
   cross-fertilisation and cross-area review.

2. We aren't doing (that much) cross-area review, or at
   least don't require it.

I disagree with message 2. The message as far as I'm concerned is
that we are not doing enough cross-area review and we aren't doing it
early enough. Frankly I am shocked by how many cross-area problems
show up in documents that reach IETF Last Call or even IESG review.
I see this regularly in the Gen-ART reviews, which you can see too at
http://www.alvestrand.no/ietf/gen/art/gen-art.html

So we aren't getting *enough* cross-fertilisation today, and we're
getting it too late in the process. That doesn't conflict at all
with message 1.


Hence, I think my question still stands: Might the benefits be greater than the costs? Might we get (almost) as much cross- fertilisation anyway?

If we need to plan meetings ahead for two years, why can't we schedule once a two weeks meeting two years ahead, for an experiment? Then we would have plenty of time, i.e. another two years, to consider whether that experiment was useful at all. And if two weeks is too much, maybe have a 1.5 weeks meeting (Sunday - Tuesday) to start with? Of course, the point would be not to spread out all areas throughout the time but concentrate "horizontal" areas towards one of the ends and try to distribute the WGs in the "vertical" areas accordingly.

Taking on purpose a simplistic view, to greatest drawback seems to be that some valuable people working on different areas would need to stay at IETF meetings for about 6 weeks instead of 3 weeks. (For the experiment 4 vs. 3 weeks).

The other drawback is that people who have a narrow focus and only stay
for part of the total time would have even less cross-area exposure,
either in the corridors or by dropping into sessions out of curiosity.

I'm suspecting that a large impediment here is fear, i.e., what if we split and those other people in the other IETF screw up things there? If so, experimenting might help to overcome that.

There is a lot of overhead in proactive liaison. We know that from
an existing experiment known as ITU-T SG13.

I think Harald's idea (which could be described as "grouped interim
meetings") is a very interesting one and would actually be a fairly
low cost experiment - if a group of WG chairs got together, they
could set such a thing up as long as the ADs involved agree.

    Brian


--Pekka

On 22. september 2005 10:03 +0200 Pekka Nikander <pekka.nikander@xxxxxxxxxxxxxx> wrote:

Have we (== the IETF) ever seriously considered splitting this
organisation "horizontally"?  That is, instead of having one IETF
worrying about all from sub-IP issues to (some) applications, have
two ones, one more focusing on issues at the IP layer and directly
above and below it, and the other one on issues "above" that?



On Sep 22, 2005, at 13:28, Brian E Carpenter wrote:

Actually that has been discussed, as have the scope boundaries
at the "top" (apps) and "bottom" (sub-IP). And the diffculty always
is the need for cross-fertilization and cross-area review.

Do you think that applications protocols can be designed with
only a liaison relationship to transport, or transport with
only a liaison relationship to the network layer?



On Sep 22, 2005, at 14:04, Harald Tveit Alvestrand wrote:

Many times. The Standard Next Message is that:

- the people who have clues in both areas are very valuable to the organization, because it's exactly them who provide the "cross-area review" that we're so proud of - if we split it, these people will either go to 6 meetings a year or stop coming to one set

That said, I've argued at times that we should try some time to have at least one meeting "split" - where different groups of groups meet in different places. However, experimentation like this is significantly in conflict with the ideal of planning meetings 2 years ahead.



On Sep 23, 2005, at 1:08, Sam Hartman wrote:

The current structure does not require document be reviewed by all
areas.  I can (and sometmies do) record a no objection vote because I
believe that adequate review has taken place without me reading a
document.


On Sep 23, 2005, at 8:59, Brian E Carpenter wrote:

That's correct. IESG ballot procedure does require 2/3 of the ADs to
record either a YES or a NO OBJECTION for a standards track or BCP
document. According to
http://www.ietf.org/u/ietfchair/voting-procedures.txt :

- "No Objection" means "I would not object if this document went forward".

    examples where No Objection might be used include:
      - I read it & have no problem with it
      - I read the protocol action & trust the AD so have no problem
      - I listened to the discussion and have no problem

    This may be interpreted as "I have no clue or have no cycles",
    in that you exercise the ability to move a document forward on the
    basis of trust towards the other ADs




_______________________________________________

Ietf@xxxxxxxx
https://www1.ietf.org/mailman/listinfo/ietf

[Index of Archives]     [IETF Annoucements]     [IETF]     [IP Storage]     [Yosemite News]     [Linux SCTP]     [Linux Newbies]     [Fedora Users]