On Mon, Jan 19, 2015 at 11:55 PM, joel jaeggli <joelja(_at_)bogus(_dot_)com>
wrote:
On 1/19/15 6:55 PM, Phillip Hallam-Baker wrote:
Before we go too far with this, it just occurred to me that implicit in
the discussion to date has been the notion that an AD spends most of
their time managing WGs.
There's an enormous fraction of this activity that is devoted to
spherding documents through post-working group stages, reviewing other
documents prior to the telechat's that punctuate iesg review. and
triaging documents where there are seriously problems. Those are not
by-in-large tied up in working group management.
Absolutely. But the point I was making is that every draft has security
implications while an application draft that has routing implications
almost certainly needs slapping with something.
Sorry to sound like a broken record on this, but I think we are making a
lot of work for ourselves by not having a reference architecture. This
creates work at the WG level as each WG goes off and makes its own
interpretation of the architectural principles. Then the IESG starts with a
set of divergent documents and tries to make sense of them.
I have been trying to do write a reference architecture to describe it to
non Internet folk and I am rather surprised to find that the Internet
architecture as currently deployed can be described in a reasonably compact
and clean model[*] without resorting to calling firewalls, NAT, VPN, SDN
etc. blasphemies.
It is counterproductive to have a WG spend time deciding how to apply JSON,
etc. and then have those decisions overridden at the IESG level to get
consistency between specs. I want the consistency, just at a lower human
cost.
One approach that might help would be for the IESG to maintain a wiki with
prior feedback from ADs sorted into some form of structure. This would then
serve as precedent folk could look up. Another pathology I really dislike
is when certain individuals wait till the AD is out of the room and then
tell the WG what the IESG is going to demand (which is frequently rather
different to what they actually say). So basically the same thing as
minutes but structured somehow.
[*] About the same number of elements as the standard model of particle
physics.