On Tue, Jun 11, 2013 at 1:45 PM, Dave Crocker <dhc@xxxxxxxxxxxx> wrote:
On 6/11/2013 5:25 AM, Dave Cridland wrote:That's the problem with special-case exceptions, such as requiring less work by an august personage. It reduces to a cult of personality and it doesn't scale. For an organizational culture of the type the IETF expresses, that doesn't fit. The opinions of people IETF management positions are not supposed to automatically have more weight in determining the specifics of our specifications; they are supposed to make their case, just like everyone else.
We want understanding, of course, but I think requiring Russ to
demonstrate that by writing a paragraph or six on the finer points of
the proposal would be daft.
We try to distinguish between comment when wearing a formal IETF hat versus without a hat. So it's not the IETF Chair making the comment, it's "merely" a well-known personage.
It's easy to give special rights to such folk, such as not requiring them to offer the substance behind their statement, but it actually has a pretty insidious effect. It's gets us used to pro-forma postings; it gets us relying on a few folk to sway things; it gets us to count rather than think.
Ah, sorry, that's not what I meant - I included Russ's name purely because he was the original exemplar, not because he's special in any particular way.
I meant that requiring anyone to demonstrate understand of the draft by jumping through hoops would, ipso facto, require them to jump through hoops.
Pete Resnick has been working on a careful formulation of what the IETF means when it talks about 'rough consensus'. My own interpretation of what he's developing -- and I want to stress this is me speaking, not me speaking for Pete -- is that consensus is a combination of both numbers and substance. The mere fact that "almost everyone" is in favor of something can't be enough. What is also required is that the arguments of objectors must have inadequately persuasive substance. One voice with a really solid concern, which withstands independent review, needs to be able to upset an overwhelming agreement.
If the politicking is from multiple organizations who all want to
implement and deploy, then I'm all in favour...
So no, the fact that the politicking is from multiple organizations needs to be insufficient.
Again, I think you're misunderstanding me - I meant (somewhat facetiously) I'm in favour of the politicking, not that I think that it should carry the day automatically. If there are unanswered objections, that should indeed count against. More generally, in the case of the XSF's small set of questions, if people answer the last call with one-word answers to those and nothing further, this gives that community sufficient information to gauge whether to advance a proposal along the standards track there - in other words, given a fairly minimal bar, any engagement meeting that bar is valuable.
That bar has to be high enough to carry more than the single bit of information Russ's note carried on its own, though, but it also needs to be low enough that it won't prove a barrier to response.
I think the XSF's questions are close to the right level (I think the XSF could, if it wanted, tweak these and improve them after this number of years); I don't think it would be very hard to find some similarly reasonable start point for the IETF.
Dave.