> On Tue, 4 Jun 2002, Lyndon Nerenberg wrote: > > While I agree with your thoughts here, the bottom line is that these > > documents talk about quality-of-implementation issues. You do not > > *have* to reference them to implement the protocol, therefore they are > > by definition not normative references.
First of all, it is OK in principle to have normative references to informational documents. The obvious example of this is a reference to any of the various hash function RFCs (MD4, MD5, SHA-1). These are informational documents but it is obvious that such a reference is normative. (The logic for these being informational is that these hashes aren't something the IETF defined. I'm not sure I agree, but that's another discussion for another day.) Having said that, I have to agree with Lyndon -- these references don't quite meet the criteria. So I don't believe they should be in the normative section. > Well, what should I do? Perhaps there should be a separate section for > so-called "informative" references which have normative content? > Remember that the old Host Requirements documents were in the same > position. Making the text that refers to these documents stronger is one thing that comes to mind. > I'm worried because we have this idiot on comp.mail.imap claiming that > IMAP is something different from the specification due to his own twisted > reading of the specification (which includes totally disregarding the > formal syntax). Granted that nothing can be done about his crusade, but > I'd like to take some sort of affirmative measures to prevent it from > happening by innocent mistake. Believe me, I'm all too familiar with this phenomenon -- it has happened to me quite a few times with the MIME RFCs. The number of idiots out there is simply amazing. While I agree that in general there's nothing you can do, especially if people persist in not following the formal syntax, I have on a couple of occasions very reluctantly changed the text to make it crystal clear that some wrongminded notion is indeed an wrongminded notion. Now, I'm not saying you should do this in this case, and I am certainly not saying we should hold up this specification to address whatever bogus claim has been raised in this case. All I'm saying is that distasteful as it may be to cater to ignorance and stupidity, it may make sense to do so on occasion. Ned