On Sep 6, 2011, at 4:33 PM, Ted Hardie wrote: > The IESG has been working to the assumption that Proposed Standards will be > widely deployed into all environments for a long time. That may well be an > appropriate response to the deployment practice (heck, if "the internet runs > on internet drafts" we're lucky that we don't have an IESG review step before > i-d publication). But if the result of this exercise is that the bar for PS > stays as-is and the bar for the second stage merges, we will retain what is a > functionally a one-stage standards process. We can certainly live with that > (we live with it now), but it means we are changing out a standard that > doesn't accurately reflect what we do now for one that doesn't accurately > reflect what we will do.
Agreed. That has been my primary puzzlement with this entire discussion. The problem that "cycling at Proposed" solves is "what if I need to change the technology in some way". "Foo updates bar" does the same. Generally speaking, we don't get a rewrite until A has been updated by B1, B2, and B3, B2 has been obsoleted by C, and C has been updated by D. What Draft Standard was supposed to fix was a raft of testing coupled with that rewrite of the specification that also removed cruft that wasn't used. Frankly, the only thing I ever figured out that "Full Standard" was useful for was "obsolete" (he ducks). I wonder if we would be better off discarding the concept of layers of standards, call PS "Standards track", and instead specify a way to report interoperability tests. If we have a document A that has been updated by B and someone has tested several implementations of A+B, could they say "I tested A+B in this configuration, which used features A.1, A.2, A.3, A.5, and B.1, with these results", without forcing the rewrite or the major conniption fits that DS involves. Folks in fact do interoperability tests with some regularity. They do them for equipment they want to buy, they do bake-offs, and they do other things. _______________________________________________ Ietf mailing list [email protected] https://www.ietf.org/mailman/listinfo/ietf
