[sc34wg3] revised draft Reference Model document N0298

Steven R. Newcomb sc34wg3@isotopicmaps.org
17 Apr 2002 12:15:25 -0500


"Ann M Wrightson" <ann.wrightson@alphaxml.com> writes:

> Your points below are all pertinent. Therefore,
> subjects are not absolute, immutable referents, even
> though it is often useful to forget that when working
> within a particular community or subject-area (this
> is what I was getting at in my previous mail in the
> PSIs thread - and see example below)).  It worries me
> that the rhetoric of the dRM and (much of the)
> surrounding discussion leads me to believe that
> subjects are somehow considered absolute a priori,
> and that consequently the RM does not appear to to be
> designed to support an AM designed to take into
> account subjects identified from different
> perspectives, other than as sub-perspectives within a
> single "God's-eye-view".

Eeeek!  I certainly hope not.  We must never confuse
a topic map with the truth.  The truth is unspeakable --
and topic maps don't make it one bit more speakable than
it has ever been without them.

> I have no problem with absolute subjects as a design
> decision - I guess I'm objecting to their appearing
> to be a religion.

I think I understand you perfectly.

(And, having said that, it seems to me that any
statement that one person perfectly understands another
is the very height of hubris.  Upon further reflection,
your remarks can be seen as an worthy effort to draw
attention to the implicit hubris, not only of the draft
Reference Model, but also of *any* technique for
communication between human beings.  All communication
is art, really.  This makes me think of topic maps as
being somehow like iambic pentameter, or
sonata-allegro, or the 1/2-hour situation comedy --
just another formal framework within which we can ply
the art of communication.)

> I'll think a bit more about this - because it v. much
> affects what is an appropriate underlying (formal)
> semantics for the RM - and BTW, I think that what we
> have in the RM is not "assertions" but "items of
> information". The difference is that assertions
> usually have an associated logic of a
> truth-functional kind - which is probably not
> appropriate in this case.

I see truth-functionality, or at least internal
consistency, as an optional feature of an Application.
(A Very Important Feature, in the case of Applications
designed to support trust.)

> An example regarding identification of subjects: if I
> have a bibliographic language describing "works", and
> another bibliographic language describing
> "documents", then using apparent identity of subjects
> (eg William Shakespeare; "Macbeth") to combine
> information expressed in the two languages can yield
> an unusable mess of misinformation (- see Svenonius's
> book (Intellectual Foundations of Knowledge
> Organization) for a thorough explanation - which does
> not depend on topic maps).

I'm not sure that Svenonius's reservations about the
collation of relations around diversely-viewed subjects
are necessarily relevant to the topic maps paradigm.
The result of merging topic maps never *has* to be an
unusable mess, because the relationships contributed by
different topic maps -- in accordance with different
world views and different ideas about the art of
bibliography -- can always be hidden or revealed to any
desired degree.

> You could disambiguate them in a topic map by having
> differentiated PSIs built using the underlying
> theories of the two bibliographic languages - but you
> need specialized knowledge in order to appreciate the
> need to do so, and many of the subjects involved
> apparently really are identical ("just are") under a
> less-bibliographically-educated perspective.

Sometimes a naive view is a useful view.  To ignore
certain kinds of distinctions is sometimes exactly
what's needed.  If I simply want to know what everybody
says about Macbeth, then I don't care who thinks of it
as a document, and who thinks of it as a work.  The
real question is whether the distinction will be lost
by combining everything said by everybody in a
convenient-to-use, fully collated form.  If we use the
topic maps paradigm appropriately and well, the answer
to that question is, "No.  All distinctions are
preserved.  You have the option of ignoring them or
paying attention to them.  What do you want to do?"

-- 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