So, the answer is, if you want to do it the IETF way, then just use
the WG working rules, and take the results to the IETF when you are
done, if the results look good enough.
No, that would be sheer lunacy.
Hmmm. Seems to me that Stef speaks sense.
No, because what Stef advocated would create an unrealistic expectation
that an external group's work would be published and/or endorsed by IETF.
It's not that IETF cannot possibly endorse such work, it's that the vast
majority of work done under those conditions does not merit an IETF
endorsement, and encouraging other groups to operate in this manner
would require that IETF spend valuable time and energy trying to
review this work. This would distract IETF from more fruitful activities.
Other groups are quite free to develop their own reputations and to publish
their own work using their own mechanisms. For IETF to be compelled to lend
its reputation to other groups' work, or its energies reviewing arbitrary
work from other groups, really is sheer lunacy.
- either doesn't consider the problem from enough different points
of view, or
Then they "request for comment" and get feedback from others interested in
the same problem.
They're quite free to do so without involving IETF.
- suffers from a lack of technical competence, or
Assuming automatically that someone from outside the IETF is incompetent is
just as arrogant as assuming that someone inside the IETF is competent.
Take a stress pill and carefully reread what I wrote. I didn't say that
work outside the IETF inherently suffered from a lack of competence,
I said that the majority of work submitted from outside groups for IETF
approval tends to suffer from one or more of several common problems, one of
which is a lack of competence. The statements are quite different.
There are other competent groups out there, but they're not routinely asking
IETF to approve work which they've previously done in isolation to IETF.
To put it another way, other groups that have established their own competence
don't need to ask IETF to lend its name to their work.
Even the work inside the IETF sometimes suffers from a lack of technical
competence.
True enough. But we accept responsibility for our own incompetence.
We're not trying to get other groups to endorse it.
- was actually developed in a (semi-)closed environment in order
to favor certain stakeholders over others
That does not preclude the work from containing the seeds of good
ideas.
No, but it often does preclude it from being adopted by IETF without major
changes. If the external group's purpose was to prototype and demonstrate
proof-of-concept, and it's willing to allow significant changes before
standardization and deployment, the fact that the work was done prior
to bringing it to IETF can be a good thing. OTOH if the external group
was trying to standardize a protocol and promote deployment of that protocol,
doing the work elsewhere and trying to fix it in IETF is usually a huge mess.
It's one thing to say that other groups would do well to use certain
parts of the IETF process, quite another to say that IETF should
endorse the work of other groups.
The IETF should endorse good work no matter where it comes from.
The IETF exists to do work, not to endorse work that came from elsewhere.
Expecting the IETF to pass any judgement on work done elsewhere detracts
from IETF's purpose and amounts to a denial-of-service attack on IETF.
The "protocol by WG committee" approach espoused by the current IETF does not
always produce good work. It is the good ideas of a few which get adopted
by the committee that makes a good protocol.
I agree. But if the folks develop those ideas in too much isolation
(as is often the case) it's not likely to be a good protocol.
Yeah, Mo, a meritocracy, and people with merit can come from anywhere.
So let 'em build their protocol, whatever it is, and bring it to the
IETF. The problems with a really bad protocol can be extremely educational
and entertaining. The elegance of a really good protocol can be extremely
educational and entertaining. I don't see how we can lose.
Think of it this way:
We'll learn something useful from almost every good protocol we review, but
we'll only learn something useful from the first two or three bad protocols
we review. After that it's a complete waste of time. It's not even
entertaining.
And if we encourage folks to design protocols outside of IETF and then
submit them, most of those submissions will be bad. Because it's far easier to
design a bad protocol than a good one, and because the people with sufficient
clues
to design a good protocol (though they do exist outside of IETF) are far rarer
than the people capable of designing a bad one.
Keith