RE: Meeting Minutes for the SMIng WG at the 49th IETF...

["Andrea Westerinen" <andreaw@cisco.com>]

Perhaps we should avoid the use of the word "partial".  My memory of this
discussion was that some constructs of a class are required and others are
optional/best effort (and this is clearly labelled).  Whether in the
"independent of everything" info model, or the protocol/encoding specific
implementations of the data models - optional versus required means the
same.

This is different than inheritance, and also different than saying that you
have a "partial" implementation.  If a construct cannot reasonably be
instantiated and provided, then it should not be in an info OR a data model.
However, if it is reasonable, useful and we want to start standardizing on
it, then it is a great candidate.

Andrea

-----Original Message-----
From: owner-sming@ops.ietf.org [mailto:owner-sming@ops.ietf.org]On
Behalf Of John Strassner
Sent: Saturday, December 30, 2000 10:41 AM
To: Durham, David; 'John Strassner'; rpresuhn-lists@dorothy.bmc.com;
sming@ops.ietf.org
Subject: Re: Meeting Minutes for the SMIng WG at the 49th IETF...


The point I was trying to make was that in the design of the
information model, a partial anything is always wrong.
However, when you map this into a data model, be it SNMP or
directory or whatever, partially implementing a class is
very commonplace.

regards,
John
----- Original Message -----
From: "Durham, David" <david.durham@intel.com>
To: "'John Strassner'" <johns@cisco.com>;
<rpresuhn-lists@dorothy.bmc.com>; <sming@ops.ietf.org>
Sent: Friday, December 29, 2000 9:18 PM
Subject: RE: Meeting Minutes for the SMIng WG at the 49th
IETF...


Hummm, the term "partial inheritance" was invented at the
sming WG meeting.
It would be more correct to say sming allows partial
implements of classes.
This seems no different than CIM, which allows optional
properties (which
don't have to be used as I am told over and over again when
I complain about
certain CIM properties :-) ). So, this concept is only
relevant to the
mapping to say, SNMP... the class structure itself does not
support "partial
inheritance" in sming. Juergen, pls correct me if I got it
wrong.
-Dave

> -----Original Message-----
> From: John Strassner [mailto:jstrassn@cisco.com]
> Sent: Friday, December 29, 2000 1:51 PM
> To: rpresuhn-lists@dorothy.bmc.com; sming@ops.ietf.org
> Subject: Re: Meeting Minutes for the SMIng WG at the 49th
IETF...
>
>
> I agree with Randy. I've never heard of "partial"
> inheritance (before the SMIng meeting, that is). It seems
> very wrong, as inheriting only some, but not all,
> attributes, methods, or relationships of a class doesn't
> make sense. If you want to inherit just some attributes,
> consider putting them in a capsule instead. For anyone not
> familiar with this terminology, it is roughly equivalent
to
> an aux class in the directory world. The key point is that
a
> capsule is an object, but not a class.
>
> regards,
> John
> ----- Original Message -----
> From: <rpresuhn-lists@dorothy.bmc.com>
> To: <sming@ops.ietf.org>
> Sent: Friday, December 29, 2000 11:24 AM
> Subject: RE: Meeting Minutes for the SMIng WG at the 49th
> IETF...
>
>
> > Hi -
> >
> > > Message-ID:
> <10C8636AE359D4119118009027AE998704F39C62@FMSMSX34>
> > > From: "Durham, David" <david.durham@intel.com>
> > > To: "'sming@ops.ietf.org'" <sming@ops.ietf.org>
> > > Subject: Meeting Minutes for the SMIng WG at the 49th
> IETF...
> > > Date: Thu, 28 Dec 2000 22:12:30 -0800
> > ..
> >
> > I think some bits got dropped:
> >
> > > Randy: Partial implementation of a class seems odd.
> There may be a semantic
> > > for the class that implies all the attributes are to
be
> used. Maybe this
> > > comes from the CMIP world. CMIP allows partial
> inheritance as a partial
> > > implementation of a class.
> > ..
> >
> > GDMO (CMIP's object definition language) supports
> multiple,
> > not partial, inheritance.  My point was that if
attributes
> > have been bundled in a class, there must be some
> underlying
> > semantic that motivates this bundling.  To then
"partially
> > inherit" from that class risks breaking whatever
semantic
> > it was that led to the creation of the class in the
first
> > place.
> >
> > GDMO's way of handling multiple inheritance and packages
> > allows one to attain the same objectives without
> bastardizing
> > OO modeling with the kind of partial inheritance that's
> being
> > introduced here.
> >
> >  -------------------------------------------------------
> >  Randy Presuhn           randy_presuhn@bmc.com
> >  Voice: +1 408 546-1006  BMC Software, Inc.  1-3141
> >  Fax:   +1 408 965-0359  2141 North First Street
> >  http://www.bmc.com/     San José, California 95131  USA
> >  -------------------------------------------------------
> >  My opinions and BMC's are independent variables.
> >  -------------------------------------------------------
> >
> >
>
>
>