[sc34wg3] Scope in the dRM

[Steven R. Newcomb]

Steve Pepper <pepper@ontopia.net> writes:

> Judging by the first diagram in the draft Reference
> Model document posted by Steve and Michel, scope is
> no longer part of the Reference Model. This surprised
> me because I had the impression as recently as three
> weeks ago (at KT2002 in Seattle) that it *was* part
> of the RM...

> [I would how can something as fundamental as this can
> change so quickly -- and when can we consider the
> reference model to be reasonably stable from the
> authors' point of view? Sigh.]

> Having now read Steve's comments on Graham's paper on
> scope, in particular the following:

> >I want to promote the whole scoping facility up and out
> >of the Reference Model and into the SAM.  Here's why:
> >...

> I realise that scope really is now gone from the
> Reference Model. Obviously this is something that
> needs to be discussed seriously. Is everybody happy
> with this?

> It also raises the question whether Steve and Michel
> consider the document that they just posted to be a
> full and complete exposition of the Reference
> Model. Is there anything at all that you haven't told
> us about yet, or can we assume that this is it?

The paper is a survey of the draft Reference Model, but
it's not the Reference Model itself, so it is not a
full and complete exposition.  In our paper, we tried
at least to mention everything that's now planned for
the Reference Model, but not all details are discussed
in the paper.

For example, the paper doesn't enumerate any of the
requirements governing the description of Applications,
such as the fact that the "node demanders" in
Application-specific syntaxes need to be identified as
such in order to permit assertions to be made about the
nodes that they demand.

Also, the whole question of subject indication and
discovery must be more thoroughly discussed in the
Reference Model.  For example, applications need to
distinguish between the kinds of assertions that impact
the subjects of certain of their role players, and
those that don't.

> Final question: What do you consider to be the
> relationship between PMTM4 and the draft Reference
> Model? Do you intend to continue to maintain and
> further develop PMTM4, or do you consider it to have
> been superceded?

I believe PMTM4, having taught us that its basic
approach was elegant and workable but that there were
still some flaws in it, should be superceded by a model
that benefits from those lessons.  The draft Reference
Model reflects all the lessons we know about.  A
surprisingly large number of people, including you,
have contributed key ideas to the Reference Model, and
some of them were elicited in reaction to PMTM4.

I hope and believe the developers of the Standard
Application will consider adopting some key ideas from
PMTM4, but I do not believe that PMTM4 should be
considered to impose any prior restraints on the design
of the Standard Application.  There are also some ideas
in PMTM4 that I would *not* want the SAM to adopt,
because they would conflict with the draft Reference
Model.

Which brings me back to your comment about scope.  I
was dismayed and surprised to discover that we could
not justify including scope in the Reference Model,
but, since then, I've managed to grow comfortable with
the idea that scope is really a SAM issue.  (It took me
a while to get used to it.  I still believe scope is
extremely important for topic maps, for several
reasons.)  By ruthlessly discarding everything that
Applications might want to change, and that really
weren't needed in order to guarantee predictable
mergability, self-description, etc., we have left
ourselves in a position where we have very good
requirements-driven arguments for the presence in the
draft Reference Model of every single thing that still
remains in it.

-- Steve

Steven R. Newcomb, Consultant
srn@coolheads.com

Coolheads Consulting
http://www.coolheads.com

voice: +1 972 359 8160
fax:   +1 972 359 0270

1527 Northaven Drive
Allen, Texas 75002-1648 USA