Re: [newtrk] draft-rousskov-newtrk-id-state-00

[Alex Rousskov <[e-mail]>]

On Tue, 6 Apr 2004, Spencer Dawkins wrote:

> We're working on our next rev of WGS(A), and you're helping me focus
> on some things I hadn't thought about (sufficiently).

I hoped for kind of the opposite effect, actually :-). I hope WGS(A)
can simply point to or, at least, reuse large portions of the proposed
declaration mechanism instead of "thinking" about these matters.

> The problems I have with "putting it on the web page" are (1)
> IETF-hosted web pages are program-generated, so we need custom
> software development for the secretariat to make this happen, and
> (2) lots of groups are using "additional web pages" now to include
> material that's not carried on the WG home page, but there's no
> consistency for these pages.

These are two side-effects of the same problem, IMO. Official WG web
space should be managed by the WG, not secretariat. I suspect there is
existing low-maintenance software that can be used to make this happen
without increasing secretariat burden. In fact, I believe that if this
is done right, the secretariate long-term burden will be much less
(and the WG burden will be higher).

This seems to be unrelated to the state declaration itself. It only
relates to the authoritative location for the declaration (draft
versus page).

> As Harald says, draft version numbers are cheap, so maybe issuing
> more draft versions is the best we can do.

Indeed, draft revision numbers cost nothing. It is the publication
delay that costs a lot. It is like communicating via snail mail: it
"works", but precludes some very useful kinds of information exchange.

> - I agree with Larry about having an enumerated set of states - maybe
> we could figure this set out, but I don't think we know them now (at a
> consensus level).

There is no enumerated set of states. There is only an enumerated set
of standard states. WGs can add their own states, not necessarily
enumerate-able.

I will try to polish the draft to make the above assertion clear. It
looks like it is a common misconception about the mechanism!

> - I like the idea of including dates like "next revision scheduled
> for", or "reviews solicited by", if people will use them.

Do you think this should be formalized for automated extraction as
well? Currently, you can include such dates in states or comments, but
no software will be able to understand them without AI.

Can we address 90% of related needs if we introduce one formal
mechanism: the state expiration date? That is, every state will have
an optional expiration date after which the WG expects the state to be
different. Should the expiration date apply to each state (for
flexibility) or to the entire declaration (for simplicity)?

On the other hand, this sounds like patching a side-effect of the
current draft expiration mechanism (which is too rigid to make
practical sense). So again, this could be out of scope, and a
different draft would be needed to solve the problem by changing
expiration semantics and allowing WGs to set meaningful draft
expiration times.

> - WGS will require responsive WG chairs (in order to have any
> effect), and this draft seems to require them to be even more
> responsive (in order to have any effect), because the necessary
> sampling rate of state changes is probably higher.

I would state it differently. Both drafts increase WG load if the WG
wants to benefit from them and do not increase WG load otherwise. The
Chair involvement is necessary at some stages, but most of the work in
determining state lies within the WG itself, as usual.

> One meta-comment - Alex's proposal would make the world a lot better
> place for people who aren't up-to-their-armpits in a working group
> trying to figure out what's going on, and Larry's suggested
> additions to a working group web page would also. I don't know if
> we're chartered to think about this or not (Scott?), but I hope
> somebody is... the level of obscurity for our work, for people who
> don't subscribe to a working group's mailing list, is pretty high.

Yes, the state declaration draft was intended primarily for folks not
intimately familiar with the draft (inside or outside the WG). It
makes little sense to declare the state of the draft to those who
already know what the state is :-). Said that, I hope the mechanism
will have a positive _stimulating_ effect on authors or draft "teams"
as well.

Thank you,

Alex.
.
newtrk resources:_____________________________________________________
web user interface: http://darkwing.uoregon.edu/~llynch/newtrk.html
mhonarc archive: http://darkwing.uoregon.edu/~llynch/newtrk/index.html