Re: COPS-PR-SPPI support with XML-SMI 2.1 (fwd)

[Michael Kirkham <mikek@muonics.com>]

I thought the rap folks might have some thoughts on this.  Some of you are
probably aware of the XML Schema I put together for SMI (MIB) module
mappings with feedback from the smixml list, but I'm not sure how much
overlap there is here iwth that list.  It's recently been extended to
allow for COPS-PR-SPPI (PIB) module mappings as well (with a couple of
caveats below) and I would like to solicit feedback.  The URL for the
schema can be found below.  Thanks.

---------- Forwarded message ----------
Date: Fri, 4 Jun 2004 16:12:47 -0700 (PDT)
From: Michael Kirkham <mikek@muonics.com>
To: smixml@ibr.cs.tu-bs.de
Subject: Re: COPS-PR-SPPI support with XML-SMI 2.1


In PIB modules, the OBJECT-TYPE's PIB-ACCESS (analogous to MAX-ACCESS) is
only specified in table objects and is not present for the entry or
columnar objects (actually they have different terminology, but the
structure is the same in the module definitions).  In that last rev, the
<access> element of the ObjectType definition was still required in all
cases.

I am wondering if anyone has a preference for this: whether it should be
made optional or if it should continue to be required but have a special
value like "not-applicable" or something.  If it is made optional, should
it be considered to have a default value (e.g. "not-accessible") for MIB
OBJECT-TYPEs, or would other implementations prefer to just generate an
error of the element is missing?

On Thu, 13 May 2004, Michael Kirkham wrote:

> Date: Thu, 13 May 2004 08:21:38 -0700 (PDT)
> From: Michael Kirkham <mikek@muonics.com>
> To: smixml@ibr.cs.tu-bs.de
> Subject: COPS-PR-SPPI support with XML-SMI 2.1
>
>
> With the addition of COPS-PR-SPPI support to [...], and SMIng dead,
> I've made changes to the schema to remove the incomplete SMIng "reserved"
> stuff and add in support for the missing parts of COPS-PR-SPPI.  This is
> currently a preliminary version, but I thought I would solicit comments if
> anyone following the list is familiar with COPS-PR-SPPI.
>
> I think I got everything, but I'll have to double-check once I've started
> implementing the conversion in [...].  Offhand, the one part that's a
> little less than satisfactory and could use some more restructuring is the
> rules for which of PIB-INDEX/AUGMENTS/EXTENDS/etc. are present.  At the
> moment the schema syntax technically permits invalid pairings.  (PIB
> OBJECT-TYPEs can have both PIB-INDEX and INDEX, but can't have PIB-INDEX
> with AUGMENTS or EXTENDS, while SMI OBJECT-TYPEs of course can have only
> INDEX or EXTENDS).
>
> Anyway, this preliminary 2.1 version is available at:
>
> http://www.muonics.com/XMLSMI/Schema/XML-SMI-2.1.xsd
>
> Other than the version number, MIB modules converted to XML according to
> this schema should remain the same.
>
> --
> Michael Kirkham
> www.muonics.com
>
>

--
Michael Kirkham
www.muonics.com