Joel, et al,
On 7/28/2011 11:54 PM, Joel M. Halpern wrote:
I think I understand your request taht we move beyond personal anecdotes. In
principle, I even agree with you.
However, when I try to act on that, I am stumped. I do not see any way to
evaluate say the last 2 years discusses to determine whether they were more
harmful or more helpful.
1. Pursue serious discussions that consider criteria for determining that. A
variety of obvious candidates for this list come to mind readily; for example:
* Does a review of the Discuss now produce a sense of its worth? That is,
do the rest of us, looking at the Discuss generally feel that it was clear and
substantive and even important?
* Does the Discuss obviously align with the criteria for a Discuss?
* Does the AD who held the Discuss now feel that it was productive, and why?
* Do the authors and chairs now feel that it was productive, and why?
* How much delay did the Discuss introduce? Was it timely?
* How much effort did it take to resolve the discuss?
* Did the Discuss produce a change in normative text?
* Does a review of the change in text, now, produce a sense of its worth?
...
It could well be worth trying to classify typical types of Discusses, such as:
* Curiosity questions. "I'd like to understand..." where no actual problem
is cited
* Review enforcement. Ensuring that particular points from other reviews
are resolved
* Personal preference. AD dislike of a working group choice, without a
clear basis the makes clear why the wg choice won't work or is otherwise impractical
* Fundamental problem. Assertion of a basic limitation to the
specification that renders it inherently problematic
* Scope. The Discuss appears to be out of scope for the document. For
example, a document intended for Draft is being challenged on basic issues that
were approved when it went to Proposed.
* Unclear. The meaning of the Discuss is unclear and/or the path towards
resolution cannot be understood because the Discuss offers no clear criteria for
resolution.
To illustrate this last item with examples that almost certainly have
never happened:
o If a Discuss said "needs to be written in an active voice rather than
passive", it would be clear what would resolve it.
o If a Discuss said "needs to be shorter", there's a clear direction for
resolution, but no criteria for for knowing when it is short enough.
o If a Discuss said "Algorithm X in the specification will not work for
the following reasons...; Algorithm Y is better for the following reasons" then
it is clear what the problem is, why it is a problem and exactly what will
satisfy the AD to resolve it.
...
It can also be worth looking at meta-issues. For example, do documents
typically get a Discuss? That is, is it typical to block progress for a
document? Does it make sense to have working groups proceed for many months and
then have presumably basic issues raised at the very last moment?
2. Review the datatracker histories over the last <x> years, sampling enough of
the documents to get a sense of the overall pattern. (The new datatracker now
supports provides a complete history for the handling of drafts; massive kudos
to the tools team!)
3. Talk with authors, chairs and ADs about previous Discusses.
But you asked that we go beyond that. How?
The above is meant to prime a serious discussion, but the first requirement is a
willingness to engage in that discussion and to pursue it seriously.
d/
--
Dave Crocker
Brandenburg InternetWorking
bbiw.net
_______________________________________________
Ietf mailing list
Ietf@xxxxxxxx
https://www.ietf.org/mailman/listinfo/ietf