[Top] [All Lists]

Re: IESG approval status of rfc2821bis

2008-06-09 11:04:53

On Jun 6, 2008, at 12:03 PM, Pete Resnick wrote:

On 6/6/08 at 2:36 PM -0400, Keith Moore wrote:

DISCUSS should only be used if the document has a showstopper that needs discussion. The IESG needs to get their act together on that.

I disagree with the statement on how DISCUSS should be used. I don't think that's either the intent of DISCUSS or the way that it should be interpreted.

I believe DISCUSS simply means that an AD believes some item of a document needs further discussion before it should be approved for publication and/or its intended status.

OK, but go back to my two criteria:

(a) The issue cannot be addressed by a non-blocking comment.
(b) The issue is so controversial that they cannot be dealt with by an RFC Editor note at publication time.

Trying not to comment on the present document but on the DISCUSS discussion: What we see in DISCUSS usage is, as is so often the case, an artifact of what the tools allow. If ADs take their document review role seriously, then they end up with hundreds of issues found on hundreds of documents after a few years. Email is not a good way to track these issues and whether they've been resolved. Issue trackers are the canonical way to track issues that can be resolved.

The way the IETF Issue tracker works is that if you enter a COMMENT, nobody can track whether the issue is still open, has been resolved etc. But a DISCUSS is trackable, and in this case trackable == blocking. That is why (a) is used only when the AD really does not care to track the issue and ensure that it's resolved.

As for (b), if an AD, either the sponsoring or the DISCUSSing AD, were to send a note to the RFC Editor to modify a document in a way the author did not agree with, I think we'd have a more serious problem. A DISCUSS is a better way to achieve consensus than battling document changes.

Whenever an author and the ADs involved agree to some text that resolves an issue, it doesn't matter much if it goes into an RFC Editor note or into a revised I-D. The essential thing is to get the agreement first!