Newsgroups: news.software.b
Path: utzoo!henry
From: henry@utzoo.uucp (Henry Spencer)
Subject: RFCs and errata
Message-ID: <1989Sep6.212344.27245@utzoo.uucp>
Organization: U of Toronto Zoology
References: <1989Aug31.035929.2430@utstat.uucp> <14612@bfmny0.UUCP>
Date: Wed, 6 Sep 89 21:23:44 GMT

In article <14612@bfmny0.UUCP> tneff@bfmny0.UU.NET (Tom Neff) writes:
>... "We regard this and that as an error in the RFC" is not
>the most useful attitude one could imagine...

Well, when (a) any sensible implementor would prefer to ignore a
requirement, (b) all existing ones do, and (c) there is no compelling
reason for it to be there in the first place, wouldn't you say that's a
pretty good case for considering it an error?  That is admittedly an
extreme case (Xref non-transmission).

Note also that the "cancel" issue is the *only* thing in our "rfcerrata"
document that represents an actual difference of opinion, as opposed to
a case where there is good reason to suspect the writers simply goofed.
More shortly on "cancel", I'm still catching up on backlog.
-- 
V7 /bin/mail source: 554 lines.|     Henry Spencer at U of Toronto Zoology
1989 X.400 specs: 2200+ pages. | uunet!attcan!utzoo!henry henry@zoo.toronto.edu