Message-Id: <3.0.32.19980608125229.00ad5490@pop.dial.pipex.com>
Date: Mon, 08 Jun 1998 15:16:13 +0100
To: Dave Crocker <dcrocker@brandenburg.com>, <drums@cs.utk.edu>
From: Graham Klyne <GK@ACM.ORG>
Subject: Re: Moving ABNF to Draft Status
Mime-Version: 1.0
Content-Type: text/plain; charset="us-ascii"

I'm a little puzzled why ABNF needs to be standard.  Why not BCP, for example?

It doesn't actually define any protocol behaviour, any more than (say)
references to RFC 2119 (which is also a BCP).

#g


At 00:00 04/06/98 +0200, you wrote:
>At 10:25 AM 6/3/98 +0200, Kai Henningsen wrote:
>>Looks to me as if what we have here, is basically that the rules weren't  
>>made to cover this type of standard. So applying them will have problems.
>
>I think that, at worst, there is some detail in the rules which is too
>specific to a particular type of specification, but that the concepts
>behind those rules are just fine.  That's why the "used for at least two
>specification efforts" approach seems to fit exactly what is required by
>the underlying principles, while invoking details appropriate to this
>category of specification.
>
>>What are the chances to declaring ABNF to be an exception to the rules,  
>
>And an exception this major is likely to cause many more problems than it
>will fix.  That's why I it could be reasonable to go with some other
>status, if only to expedite the resolution of the question.
>
>d/
>
>=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=
>Dave Crocker                Brandenburg Consulting        +1 408 246 8253
>dcrocker@brandenburg.com      675 Spruce Drive        (f) +1 408 273 6464
>www.brandenburg.com        Sunnyvale, CA 94086  USA
>
>
------------
Graham Klyne