On 17:55 24/06/2006, Keith Moore said:
That's not quite sufficient, because most WGs aren't proceeding
according to good engineering discipline (e.g. they're doing things
in the wrong order, like trying to define the protocol before the
problem space is understood), there's little external visibility of
what the WG is doing (so it's difficult to make timely input without
actually following the entirety of the mailing list discussion), and
often, nobody with any authority over the WG is checking to see
whether the WG is actually responding appropriately to such
comments. A more formal process is necessary.
+1
> Having some new sclerotic "pipeline" involved with the life
> blood of a
> working group sounds like a recipe for working group infarction to me.
In other words, we don't want to distract WGs with useful input ...
better that they should keep their heads in the sand for the entire
2-3 years of their existence and then produce irrelevant or even
harmful output. And that way, maybe a few influential people within
the WG can coerce the WG into producing something that favors their
employers' short-term interest even if it harms other interests or
glosses over important limitations.
+1
At 17:46 24/06/2006, Keith Moore wrote:
One reason that it's difficult to charter and complete WGs might be
that WGs have demonstrated a huge potential to do more harm than good.
+1
May be my French "cartesian" spirit, but I feel all that denotes a
lack of method. A methodology includes milestones and check points.
- someone/group proposes an area where to produce an IETF deliverable.
- a charter is drafted, discussed and approved for that deliverable (set).
I think it should be then appropriate that the WG
- starts understanding the charter
- possibly calls on its revision from this inner study
- describes what it wants to achieve and proposes a TOC of the
intended document, explainging how it will fit the charter. That part
should further on be used to document the deliverable and clarify
possible confusions.
- gets the whole thing reviewed by the community.
- works on the content of the described deliverable.
- loops back into that process when the WG discovers a flaws at a
previous level.
At each step any third party (other wg, IESG, IAB, users ..) could
check the consistency between the intent and the deliverable.
This methodology is the one we all use when developping a software
program. Or may be we are not? Or may be we have not much people
developping running code anymore.
For example, I am amazed by the total lack of WG requests for IAB
guidance. IAB should steer the IETF deliverable consistency.
jfc
I am amazed that m.
What the IETF does is to specify a program named "Internet".
_______________________________________________
Ietf@xxxxxxxx
https://www1.ietf.org/mailman/listinfo/ietf