[kitten] Proposal for tracking document reviews and skipping WGLC

[Benjamin Kaduk <[e-mail]>]

Hi all,

At our IETF 95 session, Stephen pointed out that the chairs do not
necessarily need to run a WGLC in order to make an assessment that there
is working group consensus for any given document; the WGLC procedure is a
customary way to do so but is not enshrined as a formal procedure.  Given
the [not-so-]recent discussion about document adoption, document backlog,
potentially abandoning old documents, and prioritizing upcoming work, it
seems reasonable to revisit our processes for finishing documents.  In
particular, with our WGLCs sometimes being extended to get enough comments
and not usually bringing in input from a large body of participants, as
well as the behind-the-scenes cajoling that the chairs have beein doing to
solicit document reviews, it is attractive to consider a way to be able to
move documents forward without needing a WGLC.  It seems that some of the
difficulty with the traditional WGLC process stems from many WG
participants not having regular (weekly or more frequent) time in which to
contribute, so the WGLC could stall until participants' availabilities
line up.

As an attempt to remedy the difficulty of coordinating everyone's
schedule, we propose to create a wiki page (or pages) where each document
can have a table of who has reviewed what version(s) of that document,
with a link to the review.  The actual mechanics of doing a review would
not necessarily change; mail still needs to go to the list with comments
and discussion, but this wiki page would help us (authors, chairs, and
participants) to track which documents are getting attention and which
might be ready to move on. [0] Once the chairs think that a given document
has received sufficient review, we can send a message to the list noting
our intention to move it forward, and start on the shepherd writeup.  In
some sense this would still serve as a WGLC, in that it would be the "last
call" for objections from the WG, but we would not have to block for a
period of time waiting for comments even if the document was in fact
ready; the comments would already be in, and the shepherd writeup could
proceed in parallel with asking if there are objections.

If this proposal moves forward, there is a question of where to host the
wiki page: two choices that came up so far are a github wiki or an
IETF-hosted trac wiki, but we are not tied to those two options.  My
understanding is that either one would require an account tied to that
provider in order to edit (to avoid wiki spam), so there would be some
barrier to entry in either case.  However, perhaps more people already
have github accounts than IETF trac accounts, which lends some preference
to github; indeed, other WGs are using github for document editing and
issue tracking already.  Regardless of where the wiki is hosted, a wiki
account would not be needed in order to participate in document review;
comments can always be sent to the mailing list and the chairs are able to
edit the wiki page on behalf of others.

Does this proposal seem reasonable?


Thanks,

Ben
for the kitten chairs


[0] This also serves as a way for an author to build good will by
reviewing other peoples' documents on the principle of "I reviewed yours;
please review mine".