Re: PROTO Process

[Fred Baker <fred at cisco.com>]

On May 13, 2009, at 7:17 AM, Andrew Sullivan wrote:
> The alternative explanation is that the "have to be modified" could  
> be stated slightly differently as "have to be modified to get  
> through" AD review, which is the sort of thing I sometimes hear  
> people saying. This locution suggests that people do not regard the  
> IETF and AD reviews as useful tests that improve the technical  
> quality of a document, but that instead the reviews are regarded as  
> hurdles to clear on the way to publishing an RFC. This might suggest  
> a reluctance to compromise that is not ideally matched for our  
> consensus-based approach. If that's right (and I don't know whether  
> it is), then no procedural tweaks will fix it either, because the  
> problem is still the suitability of the documents for publication as  
> RFCs unless we abandon rough consensus as a test.

Having been on both sides of this divide, I think I can attest that  
both views hold water.

As IETF chair and with Scott Bradner as one of the ADs, I noted that  
any draft that used normative language and didn't cite RFC 2119 would  
get a "discuss" on the point, and included that in my (decade-old now)  
id nits bullet list simply to avoid the predictable revision cycle. As  
an author on what is now RFC 4192, I found a particular AD to be using  
his position on the IESG as leverage to force into the document a  
statement which was both vacuous and untrue - and which the working  
group had not discussed. I asked him for text supporting his position,  
which he didn't supply, and his "discuss" was removed when he left the  
IESG. Those are both cases of "what it took to get past the IESG".

On the other hand, at least in my opinion, the original L2TP draft  
sent to the IESG was largely incomprehensible. Tom Narten's efforts  
with the authors, centered around "I'm not removing my 'discuss' until  
I can understand your document", resulted in a complete rewrite, which  
was a vast improvement. While that was an extreme case, I can think of  
many cases in which IESG comments have improved documents in real ways.

As an author, I sometimes get the feeling that ADs place discusses  
because they feel honor-bound to find something wrong with a document.  
I personally would appreciate getting those remarks during the working  
group process rather than having the AD wait for last call to place  
them. I appreciated Magnus' detailed comments on a draft I am working  
on in behave coming now rather than later, as they constitute a lot of  
additional effort - effort that was not deemed necessary when SIIT,  
which the draft updates, was first proposed. In general, I think that  
directorate review and AD review works best if it happens during  
working group discussion rather than as a remedial effort after the  
fact.