Re: Deployment Cases

Iljitsch van Beijnum wrote:

> But you missed my point, which is that stuff brought to the IETF will 
> almost certainly already have some measure of success, while stuff 
> developed within the IETF doesn't, because it's completely new. To make 
> that comparison fair, we'd have to compare work started within the IETF 
> with work started elsewhere regardless of whether it's brought to the 
> IETF later.
Maybe.  It depends what we are trying to understand.  If we are trying to 
understand which kinds of IETF activities have more success and which kind 
have less, then the sampling method you are suggesting does not apply.
If we are trying to compare the IETF's rate of success at its new initiatives, 
compared with the rates of other venues, then your suggestion makes sense. 
While I think that that is an entirely worthy consideration, I don't think it 
is nearly as urgent as trying to understand how the IETF should spend its 
limited resources.

> Apart from that, the IETF has been around for 20 years now and IP for a 
> few years longer than that. By now all the basic stuff has been 
(FWIW, "a few years longer" is actually 11.  TCP was first implemented in 1976.)


> invented, refined and reinvented a few times over. The low hanging fruit 
> is gone, what's left is generally trivial or (almost) unsolvable. I 
While that well might be true, I'm not sure there is any serious basis for 
knowing it to be true.  At a minimum, it would require mapping out some sort 
of topology of capabilities that could be pursued and showing that most of the 
terrain has already been explored. (I could imagine the exercise showing you 
are right, but think it best that we not accept your point on faith.)
A counterpoint is that we seem to see bursts of innovation every 10 years or 
so, with each one providing some massive improvements, including to existing 
capabilities.

> don't think other standards organizations of similar maturity knock new 
> protocols out of the park each and every time, either.
"each and every time" is just the sort of language that undermines this sort 
of discussion.  I don't recall anyone suggesting that a perfect track record 
is required, so citing it as a criterion is not very helpful.
d/

--

  Dave Crocker
  Brandenburg InternetWorking
  bbiw.net

_______________________________________________
Ietf mailing list
Ietf(_at_)ietf(_dot_)org
https://www1.ietf.org/mailman/listinfo/ietf