Process for closing out issues. (Please Read This Message).

Until Larry Masinter returns from his boondoggle, I guess I have to deal 
with process stuff.  So here's my take on a process to close out our
outstanding issues, while converging on exact wording.

Starting in the next few days there will be a series of messages
from the editorial group to the working group on outstanding issues.
Additionally, there will be I.D's in a number of areas
being handled independently (e.g. digest authentication, content
negotiation, and so on).

The next question is how to get W.G. review of exact wording
proposed to close out issues, without us all drowing in a flood of E-mail
on topics most w.g. members aren't concerned with or interested in.

Each issue in the issues list has been given a name, and it has
an issue owner.  The issue owner is responsible for
drafting the text (or getting it drafted with others) required 
to resolve the issue.

1) The issue owner will draft the text/changes to existing text to resolve 
the issue, and post the text changes (along with whatever explanitory material
is necessary) to the list. Issue owners should normally include the
issue name in the subject line.  In some areas, those typified
by the earlier subgroups, (e.g. caching), one
set of changes/additions/removals will claim to close out a whole
raft of issues.  If you care about an issue, you should
read the relevant messages, and if you participated in that
subgroup, you should review those messages particularly carefully.


If every issue, given the number of issues we have,
generates N mail messages, discussing this or that
sentence needs work, we'll all go nuts, particularly if we all have
to wade through each message to figure out if it is an issue you
care about.

2) So here's my suggestion: first try private e-mail to the issue
owner to resolve problems, rather than CC'ing the whole list.

If you can't resolve the problem privately, then by all means
take the issue public for public discussion.  But at least this
way wording problems won't generate a flood of mail.
Please try to keep the issue name in the subject line (or beginning
of larger messages that close out a raft of issues), so that
everyone doesn't have to wade through every message to figure out
if they care about it or not.

Next problem: how to get a feeling for concensus...

3) When the issue owner believes he has IETF "rough concensus" on an issue,
he'll repost the changes to the list with the final wording additions/changes, 
and will indicate he believes the issue has reached rough concensus.
(again, with the issue name(s) attached). 
In many circumstances, it would be useful for private mail to be
summarized with the issue so that others can see the reasoning
behind the resolution, so that we don't have to repeat the
whole discussion on subtleties of the issue on the general list.
For obvious, non controversial changes, this will result in 1 or 2 messages 
to the list to close out an issue.

4) At this point, you should have commented in private if you care about
the issue, and if it isn't resolved to your satisfaction, we
can hash it all out on the general list.  I hope, however, that
as much technical and editorial stuff as possible can be resolved without every
argument being done on the general list, and we'll be down
to disputing if the text actually reflects the concensus on the issue,
and issues that are still really in dispute.

My hope is to have another complete draft together by around 4/1,
with a final draft in mid-april.  As they become available, we'll
do something similar to the above on problems that turn up in the integrated
drafts.

If I'm all wet or you think of a better process, do let
me know.

I will be keeping issues status until Larry Masinter returns, as usual, 
in the issues list found at:

http://www.w3.org/pub/WWW/Protocols/HTTP/Issues/http-wg.html

Please refer to the issues list before raising what may
be a resolved issue, or for the status of an issue.  
I'll try to keep pointers there to messages resolving issues.
			- Jim

Received on Friday, 22 March 1996 11:05:09 UTC