Re: problems with RFC1227, RFC1238, and RFC1248

kolb@psi.com Thu, 22 August 1991 20:04 UTC

Received: from nisc.psi.net by NRI.NRI.Reston.VA.US id aa17957; 22 Aug 91 16:04 EDT
Received: by nisc.psi.net (5.61/2.1-PSINet Operations ) id AA08617; Thu, 22 Aug 91 15:41:54 -0400
Received: from psi.com by nisc.psi.net (5.61/2.1-PSINet Operations ) id AA08613; Thu, 22 Aug 91 15:41:52 -0400
Received: from localhost by psi.com (5.61/2.1-PSI/PSINet) id AA04485; Thu, 22 Aug 91 15:41:44 -0400
Message-Id: <9108221941.AA04485@psi.com>
To: Keith McCloghrie <kzm@hls.com>
Cc: snmp-wg@nisc.psi.net
Subject: Re: problems with RFC1227, RFC1238, and RFC1248
In-Reply-To: Your message of "Thu, 22 Aug 91 10:51:30 PDT." <9108221751.AA16546@nms.netman>
Date: Thu, 22 Aug 1991 15:41:42 -0400
From: kolb@psi.com
Status: O

> From: kzm@hls.com (Keith McCloghrie)

> Chris,

> > 	RFC1227-MIB (SMUX):

> > 	  macro reference "OBJECT-TYPE" is imported from mib reference
> > 	  "RFC1212".  no such mib reference exists.  the correct mib
> > 	  is reference "RFC-1212".

> Actually, RFC 1212 defines neither "RFC1212" nor "RFC-1212".  
> So, "correctness" to this level/depth is not achievable by any 
> of today's MIBs.

 my mistake...  i was looking at a version that i had corrected
 awhile back so that it was legal ASN.1 code.

 so the result of this lack of correctness is that we have taken a
 language, ASN.1, that is unreasonably difficult to parse, though
 parsable given enough time, effort, and cranial discomfort, and are
 treating it in standards documents as if it were a machine imparsable
 language.

 but MOSY parses this stuff, why can't my compiler parse it?!?  why?
 because MOSY doesn't do very much type or scope checking.  my compiler
 attempts to do very strict type and scope checking.  while testing my
 compiler, i ran every single MIB (RFC, internet-draft, and vendor MIB
 from venera) i could get my electronic hands on, through my compiler.
 they all passed through MOSY, but over a dozen wouldn't pass through
 mine without changes, some of those changes being fairly extensive.

 now i'm not trying to bash MOSY, and i am also not trying to tell everyone
 that i have a great compiler (it's still in it's infancy).  the bottom line
 is that passing a MIB through MOSY tells you very little about the actual
 correctness of the documents in terms of being legal MIB modules or even
 containing legal ASN.1 code.

 perhaps both the RFC1212 and RFC1215 documents should be reissued as legal
 ASN.1 code (iow, scoped within module reference definitions).

> > 	RFC1248-MIB (OSPF):

> > 	  value reference "ospf" is defined in terms of the value
> > 	  reference "standard-mib", which neither imported into nor
> > 	  defined in mib reference "RFC1248-MIB".

> > 	    ospf OBJECT IDENTIFIER ::= { standard-mib 13 }

> > 	  "standard-mib", i assume, really should be "mib-2", which
> > 	  will need to be imported from mib reference "RFC1213-MIB".

> RFC 1248 has been re-issued as RFC1252, correcting this problem.

 i didn't catch that...  thanks!

> Keith.

 chris