Relay-Version: version B 2.10 5/3/83; site utzoo.UUCP
Posting-Version: version B 2.10.1 6/24/83; site joevax.UUCP
Path: utzoo!watmath!clyde!burl!ulysses!gamma!pyuxww!pyuxa!bambi!joevax!leland
From: leland@joevax.UUCP (Lee Woodbury)
Newsgroups: net.music.synth
Subject: A Context-Free Grammar for MIDI?
Message-ID: <49@joevax.UUCP>
Date: Thu, 12-Sep-85 10:32:00 EDT
Article-I.D.: joevax.49
Posted: Thu Sep 12 10:32:00 1985
Date-Received: Fri, 13-Sep-85 04:38:39 EDT
Organization: Bell Communications Research (Bellcore), Morristown, NJ
Lines: 28

As a relative neophyte, I, for one, was delighted with Bob McQueer's
"MIDI Primer" posted to net.music.synth recently.  Thank you, Bob.

Reading it, however, and having a computer science background,
I couldn't help thinking that the MIDI software spec would be even 
clearer if it were presented in the style of context-free language 
specifications, in Backus-Naur Form (BNF) or some other context-free 
grammar.  [Aside: is MIDI a context-free language?]  Also, it seems 
to me that such a specification, in conjunction with
a "compiler-compiler" program like yacc, would be a useful step 
toward creating a software module which parses MIDI IN and/or generates
MIDI OUT sequences.

Is anyone out there aware of such a spec (in the public domain),
done either in the context of simply describing MIDI or
in the context of MIDI software development?  I'm not very experienced
with MIDI: are there reasons why this might NOT be feasible or
useful?  Does it strike others as a good idea?

Much obliged for any response.  Feel free to mail to me, but post
to the net if you think it's of general interest.
-- 
Lee Woodbury
Bell Communications Research (Bellcore)
435 South Street, MRE 2D-290
Morristown, NJ  07960
201-829-4535
..!{allegra,ihnp4}!bellcore!joevax!leland