Brian E Carpenter wrote:
On 2007-06-27 17:42, Michael Thomas wrote:
Brian E Carpenter wrote:
One thing that would make a significant difference would be if WGs
really took responsibility for their own quality control. Even at the
trivial level, the IESG still gets drafts that don't pass ID-nits
(but that is getting better, thanks to PROTO shepherding). But maybe
we should be pushing harder to make WGs responsible for getting
final external reviews done (and responded to). That would just be
more delegation along the path that's been followed for the last
several years.
Part of the problem with cross area review is that there's such a flood
of documents that it's pretty impossible to know unless somebody
tells/asks
you to review it that you should care.
Indeed. The current reviews such as Gen-ART and secdir reviews don't just
happen spontaneously; specific reviewers are triggered by a "dispatcher."
As I said, I'm a little skeptical about
"expert review", but for so-called experts as well as the laity in
other areas,
it might be nice to have some indexing when it touches areas which are
known to contain land mines.
We also need reviewers to look for land mines in unexpected places.
We already do this to some degree with Security Considerations, but they
often read as a "Full Disclosure" part of a contract rather than the
basis of
how operation security is intended in the draft itself.
What might be nice is to have some cross domain Cliff's Notes capturing
the jist of how this draft uses or affects other areas? So that
people in other
areas can determine more easily if something whacky might be going on
and pique their interest in reading it? Were this done sooner in the
process
rather than later, we well might be able to nip more whackiness
before it
bubbles all the way up to the IESG.
It sounds like that would be done before passing the document
on for AD review. Is that your idea?
Yes -- well before. Maybe something something along the lines of
what it builds on, might impact, might be applicable to, etc. The
problem that I see is that there a fair amount of beg, borrowing and
stealing (good) and that is often done without the keepers of those things
knowledge until pretty late in the game (bad). Take DNS for example
since it's a favorite franken-building block. Would anybody in the
namedropper crowd be likely to know if, say, the routing crowd
was using DNS in some new way? Probably not. If, however, we
had a dashboard which named out the cross-area dependencies in,
oh say, the area section or the working group page or somewhere
like that it would at _least_ give you a clue that there might be something
interesting to look at in an area you normally pay no attention to.
Mike
_______________________________________________
Ietf@xxxxxxxx
https://www1.ietf.org/mailman/listinfo/ietf