CURRENT_MEETING_REPORT_

Reported by Peter Furniss/Consultant

Minutes of the Minimal OSI Upper-Layers Working Group (THINOSI)


Status and Content of Upper-layer Cookbook

The question of whether the cookbook was a parallel specification of the
OSI protocols or a mixture of profile and implementation advice was
finalised, following earlier e-mail discussions.

It transpired that the Area Director, Dave Crocker, and several others
had originally thought that the group was specifying an alternative
protocol to provide the OSI upper-layer functions, and they had been
surprised to discover the protocol was the same, or a subset of the
standard protocols.

The group accepted Dave's view that there could be no such thing as
``re-specification'' of a protocol---there was only one defining text.
Anything which restated, without modification, what was in the original
specification was really an implementation guide.  As such it could
become an Informational RFC, but not a standards track document.

However, the cookbook also subsets the standard protocols, and in this
respect is similar to the profiles produced by the OSI ``regional
workshops''---OIW, EWOS and AOW. Such subsetting is protocol
specification, and Dave said would be suitable for standards track if
the base documents are:


   o The result of an open process.
   o Stable (for Proposed status); or published standard (for IETF
     Standard status).


Dave suggested the model of the SMI RFC, which cites the OSI ASN.1 Basic
Encoding Rules and defines a subset of them, could be taken as an
example of the style.

Peter Furniss had already produced a first attempt at separating the
profile-like aspects of the cookbook, treating the OSI base standards
and the two general profiles (Common Upper-Layer Requirements - Part 1
and Part 3) as the cited documents.  Both the ISO and regional workshop
processes meet the openness requirement.  However, although CULR Part 1
(general requirements) is stable, and about to begin draft International
Standardised Profile ballot, Part 3 is not stable and is some way from
international ballot.  It was agreed that Peter would expand the
profile, citing just the base standards and CULR-1 and reproducing the
CULR-3 restrictions.  It includes a few further restrictions that are
not currently in CULR-3.  Peter would submit this as an Internet-Draft.
Peter will also revise the cookbook again, referring to the other
document.


Charter Revision

The decisions above require revision of the charter to reflect what the
group is actually doing.  It was agreed that since the ``thinDAP'' work
has not progressed, and in any case would be an Informational RFC, it
should be dropped from the work plan.

Peter will work up a draft revision and post it to the list, (really).


Application-specific Mapping Documents

There was no clear view on the possibility of defining application-
specific mapping documents (e.g., how to use Z39.50 with the cookbook).
Peter will explain this idea to the mailing list.


Reduced-OSI

Following the realisation that many people had expected, and wanted, the
group to investigate alternative, lighter, protocols, it was agreed the
mailing list would be opened to provide an (interim) forum for
discussion of this.  Walter Lazear had identified at least 7 different
groups (mostly .gov or .mil) interested in this, and at least two others
were represented (electric power companies) or known of (civil
air-ground).  Walter had a one-page summary of this to hand out.  The
concentration would be on trying to establish what the requirements
really were---i.e., which parts of OSI function were still wanted.

This widening of the list will not be formally part of the working group
work plan.  It is just taking advantage of a mailing list that people
may have thought was doing it anyway!


Parallel Documents

   o Jim Quigley reported that CULR-3 had been revised further, and the
     OIW ULSIG were concentrating on getting the compliance/conformance
     terminology sorted out.

   o The X/Open XTI/mOSI specification is still about to be published as
     a preliminary specification.

   o Jim Quigley reported that ITU-T SG7 was planning to make the
     cookbook into a Recommendation.


Implementations

   o Terry Sullivan (Florida Center for Library Automation) released his
     ``tosi'' implementation three days previously

   o Peter has started extending the X/osi code to a more general
     thinosi implementation, with XTI/mOSI as the upper interface.  He
     hopes to have it available around the end of the year.


Next Time

The group does not expect to meet in Seattle.


Attendees

David Crocker            dcrocker@mordor.stanford.edu
Richard desJardins       desjardi@eos.nasa.gov
Peter Furniss            p.furniss@ulcc.ac.uk
Walter Lazear            lazear@gateway.mitre.org
Mark Needleman           mhn@stubbs.ucop.edu
Dan Nordell
James Quigley            Quigley@cup.hp.com
Brien Wheeler            blw@mitre.org