[sc34wg3] Semantics of subject, topic type, etc

Murray Altheim murray06 at altheim.com
Wed May 31 18:42:58 EDT 2006

Quoting Lars Marius Garshol <larsga at ontopia.net>:
> * Murray Altheim
>> In your blog entry "PSIs for Topic Maps constructs" you state:
>>  "One thing that's lacking in the current set of Topic Maps
>>   standards is defined identifiers for the Topic Maps constructs,
>>   like subject, topic, association, etc."
>> I don't follow this at all. As you know, with the publication of
>> XTM 1.0 in December of 2000 there were PSIs for all Topic Map
>> constructs.
> That's true. I forgot about that. The blog entry was written very
> quickly during the meeting, so I didn't really spend much time on the
> introductory text.
> However, core.xtm didn't have PSIs for all the constructs, only for
> topic, association, and occurrence (plus the class-instance,
> superclass-subclass, and sort/display ones). Also, it lacks a formal
> definition of the semantics, and it was really the formal semantics
> that was the purpose of this exercise (and that's probably why the
> existence of core.xtm did not occur to me).

I realize you weren't there for the XTM 1.0 discussions, but
we very deliberately decided to create PSIs for all Topic Map
constructs that we wanted to be able to distinctly reify, and
no more than that. We created two necessary "Association Templates"
(though we were unable to agree upon the necessary details
for the definition of an Association Template within the time
we had, as that was heading into schema language territory, and
certainly loath to call them as such absent a definition). What is
in core.xtm was the result of that, for better or worse, but it
is tied utterly and completely into the text of the specification.

As for formal semantics, I don't see that your blog entry is any
more formal than the prose text of the XTM 1.0 prose text. It's not
really formal semantics (at least in my understanding of the use
of "formal") until one has provided a mathematical underpinning to
the model, then tied that to human-readable documentation (human-
readable being a relative term to non-mathematicians). What XTM 1.0
did was provide a functional definition for each of its constructs,
then provide PSIs for those necessary to create Topic Map instances.
Following on after XTM went public, a mathematician did actually
write up a formal semantics for XTM but I don't have the reference
to that handy. (anybody have a reference to that? probably 2001/02?)

As a similar example, the upcoming ISO standard for XML Common Logic
will have a true "formal semantics", a prose definition, and "PSIs"
(URI identifiers) for all major language components. It will also
have a tutorial text as an adjunct to the standard. I don't bandy
about the term "formal semantics" any more myself because it's such
an abused and overused term nowadays. Hell, "semantics" should be
stricken from English for a few years to get people to calm down
and stop thinking they're doing something new and remarkable.

>> You may feel the need to create PSIs for the new documents [...]
> Yes, a decision was made by ISO to define new PSIs.

Certainly necessary.

>> Revisionist history or did you mean something else?
> Forgetfulness, actually. :)
>> Your proposed set seems merely a revision of the existing ones,   
>> updated for the
>> TMDM. Obviously a necessary and good thing.
> An updated version with semantics added, yes.

There were semantics in both ISO 13250:2000 and XTM 1.0. Unless
you're planning to provide actual formal semantics, e.g.,


then you might consider dropping the word "formal." It's just prose.


Murray Altheim <murray06 at altheim.com>                              ===  = =
http://www.altheim.com/murray/                                     = =  ===
SGML Grease Monkey, Banjo Player, Wantanabe Zen Monk               = =  = =

       In the evening
       The rice leaves in the garden
       Rustle in the autumn wind
       That blows through my reed hut.  -- Minamoto no Tsunenobu

More information about the sc34wg3 mailing list