[topicmapmail] Topic Maps, Description Logics and Semantic Networks

[Giuliano Vivanet]

Scrive Murray Altheim <murray06 at altheim.com>:

Dear Murray,
firstly please forgive my delay in answering... I have preferred to make some
research about your suggestions before this reply.

> Are you investigating/comparing surface structures or their underlying
> abstract models?

We are in an explorative stage of this analysis, but I think we will focus the
attention on both of them.

> I'm curious because a number of the technologies
> mentioned in this discussion have no abstract model that ties into any
> common framework, hence any comparison is by definition subjective.


I agree with you about the subjectivity. Moreover there are some kind of
semantic networks, often very different each other... but probably we can focus
our attention on KL-ONE.

> For those that do reference a mathematically-sound model there can of course
> be a real comparison. I would urge you to check out the work on Common
> Logic (CL) as that would at least tie RDF down to CL (according to Pat
> Hayes, who wrote the model for RDF and is involved in the CL work), and
> I believe someone *may* have already done the work to tie DL into CL,
> but if not the language of DL can be linked directly to FOL so that
> shouldn't be too much of a stretch (though of course one is a superset
> of the other; you'd be left out on a limb for anything not defined in
> FOL, which with DL is quite a bit).
> Topic Maps does not (yet, to my knowledge) make any reference to CL (or
> FOL) so this can only be added as an ontology *on top of* the existing
> Topic Maps model (effectively ignoring the TM model and simply using it
> as a graph carrier for the CL or FOL semantics via a PSI set, which is
> what I think most people do and would recommend given the absence of an
> alternative -- the two Association Templates in XTM were left deliberately
> vague since we didn't have any scope to create a mathematically-sound
> model at that time, though I think some French mathematician wrote one
> subsequently).
>
> There is a CG set of PSIs in the Ceryle release and I believe a few
> other people have done work along these lines. The CL spec is meant to
> have URI identifiers for all constructs so when that is finally
> published those would be suitable identifiers.
>
> Not sure if this is any help. (Common Logic is easily found via Google)

Your suggestions are very helpful for me. In particular I want to examine better
CL following your advices.

Thank you so much!

Giuliano


> Murray
>
> ...........................................................................
> Murray Altheim <murray07 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
>