[sc34wg3] revised draft Reference Model document N0298

[Steven R. Newcomb]

"Martin Bryan" <mtbryan@sgml.u-net.com> writes:

> Thanks for taking the time to explain your thinking
> to me. I think the key is the (somewhat incomplete)
> last sentence: "However, I can say this much with
> certainty: a single 13250 <topic> can demand the
> existence of several nodes and several assertions at
> the dRM graph level."

> I say incomplete because I thing the words "of
> different types" should be included after "several
> nodes".

Yes.

> I also think that there is another key problem in the
> overloading of the term topic, as stated in: 'OK,
> then your operative definition of "topic" is not the
> same as the draft Reference Model's operative
> definition of "topic". ' Later in the same paragraph
> you introduce the concept of a "topic node". I would
> strongly suggest that this term be adoped in the dRM
> to clearly distinguish between an identifiable node
> in the dRM and an identifiable topic in 13250.

Michel and I have been promoting a usage of the term
"topic" that is designed to help us distinguish whether
we're talking about the map or the territory.
"Subjects" are the territory, while "topics" are
Topic-Maps-defined surrogates for subjects, regardless
of whether topics are syntactic object (such as <topic>
elements) or an in-memory objects that represent nodes
(i.e., single subjects) in a graph.  Making this
distinction in this way has worked well for us in our
teaching and communicating.  However, it does tend to
lump radically different things together into the
definition of "topic", and it deliberately forces into
the background other important distinctions, such as
the many distinctions between the realm of
syntactically interchanging topic maps on the one hand,
and the realm of exactly what is meant, at the dRM
level, by instances of Topic Maps syntaxes, on the
other.

If you believe we should stop invoking the "map
vs. territory" distinction by means of the
terminological dipole of "topic" vs. "subject", then
maybe we can call nodes in graphs "subject nodes".
What do you think about that idea?

If we're religious about saying "topic node" when we
mean node, then why shouldn't we also be religious
about saying "<topic> element" when that's what we
mean?

> I'm not sure that I understand how your
> topic-basename assertion type links to the topic role
> statement in the preceding diagram.

We're about to post a brief illustrated explanation
about that.  I hope it helps!

> However, this is not the cause of my problem, which
> is to do with the restrictions on whether the same
> 13250 topic can be referenced as both an assertion
> type and a role type.

I don't think it would be right to do that.  I don't
see how a subject that is a role can be the same
subject as a subject that is an assertion type -- even
if both subjects happen to have the same name.

> I quoted the example of a topic called
> "PublishedBy". This could be an association
> type.

Sure, it could be, but *is* it?  It is the topic map
author's responsibility to have exactly one subject in
mind for each topic.  The author is not required to
make clear exactly what that subject is (indeed, to
impose such a requirement would be like requiring all
topic map authors to be great writers, illustrators,
artists, animators and film directors), but the author
*is* required to have some single unique and distinct
idea, at least in his own mind.  As delivered by an
author, a topic map should contain no topic that has
more than one subject, or the about subject of which
the author has any confusion.  (A topic can have a
subject which is the author's confusion about some set
of ideas, but any such confusion is still a single,
unique subject.)

Topic maps that do not observe the "one topic-one
subject" constraint are not reliable as maps.  They are
like maps of Texas on which there is a black dot which
is usually labeled "Dallas" and, but, from time to
time, the label of the same dot says "Houston."

> Would this be a role for an assertion, or an
> assertion type for the assertion?

It could be either, but it cannot be both, even if the
name is ambiguous.

> If the latter, then what is to stop the same topic
> from being used as an occurrence role in another
> assertion?

In dRM terms, what would stop this from happening is
the the rule that no single subject can serve as both

  * the R end of any CR arc, and

  * the P end of any AP arc,

at the same time.


-- 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