ietf
[Top] [All Lists]

Re: Content-free Last Call comments

2013-06-12 04:28:34

perhaps we should go to the source of the problem and require a phd
dissertation and defense from draft authors.

A couple of years ago I worked with someone who completed his PhD thesis on a 
topic faster than it took to publish the RFC on the same topic… that was my 
wake-up call for IETF process needing some work :-) Things have improved since 
then, though, but probably not enough.

But back to the topic. I, for one, would like to see responses on IETF last 
calls. It builds my confidence that we know enough about the topic to make an 
approval decision. Particularly when the input comes from people outside the 
working group. And I'd like to distinguish "everyone thinks this is fine" from 
"no one read the document".

And substance of the comments - be it positive or negative - is obviously 
important. But I find myself strangely between Stephen's opinion and Pete's 
opinion. Pete is of course right that blind "support" is bad and that we could 
easily guess wrong what the commenter meant, even when there is substance 
behind the support but the substance did not get explicitly articulated in the 
e-mail. But I can also see that there are cases where there's some context. 
Elwyn's mail illustrates an important example with Gen-ART reviewers. (Thank 
you by the way Elwyn and others for this great service. It is very much 
appreciated, and does improve RFC quality.) 

Anyway, we know why these reviews are being done, and what the expectation is. 
The expectation is that a generalist reads the spec and determines if it is 
understandable, tries to spot possible errors, and so on. But at the same time, 
the reviewers are unlikely to be in a position to say, for instance, that they 
have implemented the spec or plan to deploy the technology. So at the end of 
the day, if the spec is fine, it will only say "Ready", We will interpret that 
as an Internet technology generalist outside the working group having read the 
document and indicating that it has no obvious problems. But there is no other 
implication - about the reviewer needing the work for himself or herself, for 
instance.

In Russ' case I took the message to mean that he reviewed it as an expert on 
the technology. It would probably have helped if he said whether he only 
reviewed it for correctness or if he was also making a statement about the 
technology being needed in his opinion. I would have appreciated his opinion on 
it. That being said, it feels kind of odd to explicitly say things if you have 
nothing to say about the matter. "I have read the document and it seems fine, 
but I am not implementing it and I don't know if it addresses a real need in 
the Internet." We tend to assume that if the commenter is not saying these 
things, it is because he has no statement about his implementation, for 
instance.

Jari