Re: WG Last Call draft-ietf-sming-reqs-04.txt

[Juergen Schoenwaelder <schoenw@ibr.cs.tu-bs.de>]

>>>>> Ayers, Mike writes:

Mike> We cannot specify our way out of bad implementations - "what"
Mike> and "how" are different questions.  This requirement is also
Mike> meaningless (of COURSE it must be machine readable, or else
Mike> parsers can't exist) and should be lost.

If the "what" is not precisely defined, then it is much harder for
implementors to get the "how" right. RFCs 2578-2580 do not precisely
define the syntax of SMIv2. It relies on a ASN.1 document which is
(a) is not easy to obtain - even for money and (b) which is itself
hard to read and understand.

Mike> Frankly, an ABNF grammar is more likely to lead to bad parsers
Mike> ("always wanted to build me a parser - here'n's the spec!") than
Mike> improve the quality of existing parsers.

Interesting. You should talk about this with compiler folks. They
probably can convince you of the value of a *BNF grammars for
specifying languages.

/js