Path: utzoo!utgpu!news-server.csri.toronto.edu!cs.utexas.edu!uunet!munnari.oz.au!cs.mu.OZ.AU!cs.mu.oz.au!kre From: kre@cs.mu.oz.au (Robert Elz) Newsgroups: news.software.nntp Subject: Re: NNTPD hates Message-IDs with TWO '@'s in them. (BIG log file attached to this posting) Message-ID: Date: 19 Jun 91 13:06:30 GMT References: <1991Jun16.063341.13609@kithrup.COM> <6021@gazette.bcm.tmc.edu> <1991Jun17.193252.28272@zoo.toronto.edu> <5127@lib.tmc.edu> Sender: news@cs.mu.OZ.AU Organization: Comp Sci, University of Melbourne, Australia Lines: 37 sob@lib.tmc.edu (Stan Barber) writes: >Okey. Let's look at the following article. >From: eggert@twinsun.com (Paul Eggert) >Subject: Recently observed nonconforming Message-IDs (discussion) And then shows lots of bugus message id's again, with explanations, including these ... > 32 The local-part may not end with `.'. > 22 The local-part may not begin with `.'. > 17 Two adjacent unquoted `.'s may not appear in a Message-ID. Since this is the second time (third if you count the later reposting of this article without comment, and for no apparent reason) I'm bewildered as to why no-one has commented on these. Exactly why are these things supposed to be invalid. The example given for the latter had the two adjacent dots in the domain part, which certainly is illegal, but that's not the explanation given. What's supposed to be so magic about dots in the local part that anyone would want to limit their use? I would submit that <...@cs.mu.oz.au> is a perfectly good Message-ID - if not likely to be a very useful one to use, but only because it will start getting absurdly long if the technique to make unique variants was just to add more dots. This breaches all three of the "rules" above, but nothing in any of rfcs 822 1123 or 1036. kre