[topicmapmail] Sample ontology/datasets
Sebastian Samaruga
I'm a developer (still) trying to get into semantics. I like very much
TMs, but there also need to use RDF for some purposes. I'm actually
trying to build an abstract ontology (it is really the model for an
index) and I'm needing sample data in any format (TMs, RDF, SQL, CSV,
etc) to try out the mappings and if I can go from/to the data formats
and my actual entities (import/export). Any sources information will
be appreciated. Regards,
Sebastián Samaruga
http://cognescent.blogspot.com
_______________________________________________
topicmapmail mailing list
topicm...@infoloom.com
http://www.infoloom.com/mailman/listinfo/topicmapmail
Murray Altheim
> Hi,
>
> I'm a developer (still) trying to get into semantics. I like very much
> TMs, but there also need to use RDF for some purposes. I'm actually
> trying to build an abstract ontology (it is really the model for an
> index) and I'm needing sample data in any format (TMs, RDF, SQL, CSV,
> etc) to try out the mappings and if I can go from/to the data formats
> and my actual entities (import/export). Any sources information will
> be appreciated. Regards,
Hi Sebastián,
Like it or not, but a "source" always comes with baggage.
The idea of converting between sources that come with differing
ontological commitments would be considered by anyone with a
mathematical bent as distinctly an erroneous proposition. You
really can't *simply* transliterate between logical systems
without resolving those differences, though it's commonly done
by programmers who care little for the semantics. The fact that
two different systems both have a similarly-named predicate is
not the basis of a merge (see below). This kind of epistemological
violence is unfortunately done daily by many users of RDF, and
a lot of boy programmers.
So (heavily loading the question), are you at all concerned with
the ontological commitments [1] of your tools? I.e., are you willing
and/or interested in being tied into First Order Logic (FOL) [2],
Description Logics [3][4], Conceptual Graphs [5][6], etc.; are
you willing to develop or adapt your own set of commitments (base
ontology); or are issues around epistemology not all that important
to you? Perhaps a particular kind of logic is your interest
(fuzzy [7], modal [8][9], topological/epistemic [10], or perhaps
mereotopology[11]?)
There's a lot of research going on and a lot of new ideas floated
in just the past decade, hard to keep up... there's also graph/
network/set theory, all manner of representations. I'd recommend
John Sowa's book on Knowledge Representation [12] but it'd hard
to get a copy (you have to contact him directly, though I'm sure
he'd be pleased to sell you one, I think he has a pile in his
basement... ahh, I see some on Amazon..).
You write you'd like to "get into semantics" so I'll assume the
epistemological basis of your ontology might be important to you.
If your primary interest is in FOL, or fairly strictly-defined logics,
or you're interested in machine inferencing, RDF and OWL would be
quite suitable. If that's the case, you'd be using Description Logics
(DL) as that's the basis of OWL. RDF has its own formal model theory
[13] but it's (to my understanding) completely compatible with OWL.
This might be important, if you play by the rules -- you'd be an
extremely rare individual if you did. In any case it'd be good to
read up on DL if you're working in that area.
Or perhaps look into Cyc/OpenCyc [14][15], which is an existing,
fully-developed ontology with lots of tools and documentation
available. Very cool stuff, and certainly model-able in Topic Maps
(I've done it many years back). Cyc provides a truly enormous data
set, lots of ontology stuff to get your feet wet.
Or if perhaps you're more interested in modeling language use, i.e.,
the way humans make statements, something like Conceptual Graphs
might be suited at the low end (before we get into serious
computational linguistics). This is more about creating sentences
in the form of assertions (triples of speech, so to speak). This
is close to my primary interest, e.g., [16][17].
One of the benefits of Topic Maps is that there are no required
semantics native to the environment (apart from the two relations
used by Topic Maps itself, which you probably should ignore in
your own model). In Topic Maps you must provide your own logic,
which is a strength (flexibility), not a weakness (constraint).
Another benefit of Topic Maps is that it *is* possible to create
mappings between different systems -- indeed, it's probably the
only environment both suited to and designed for this very thing.
It's hard to get across to people this feature, and the profound
importance of it. Topic Maps is probably the *only* "safe" way to
model how two different systems/peoples/political groups/etc.
look at what *might* be the same idea, to be able to safely model
contextually-bound concepts. As Steve Newcomb has put it, Topic
Maps enable "Global Knowledge Interchange" [18][19].
If you're going to read *any* of the links I've provided, please
start with [19], then [17]. I'll assume you've already seen 'The
TAO of Topic Maps' [20]...
Have fun,
Murray
------------
[1] http://en.wikipedia.org/wiki/Ontological_commitment
[2] http://en.wikipedia.org/wiki/First-order_logic
[3] http://en.wikipedia.org/wiki/Description_logic
[4] http://dl.kr.org/
[5] http://en.wikipedia.org/wiki/Conceptual_graph
[6] http://conceptualstructures.org/
[7] http://en.wikipedia.org/wiki/Fuzzy_logic
[8] http://en.wikipedia.org/wiki/Modal_logic
[9] http://www-formal.stanford.edu/jmc/mcchay69/node22.html
[10] Dynamic Formal Epistemology, Patrick Girard, Springer 2011
http://www.springer.com/philosophy/book/978-94-007-0073-4
[11] http://en.wikipedia.org/wiki/Mereotopology
also check out Barry Smith's pages, which provide a broad outline:
* Logic and Formal Ontology
http://ontology.buffalo.edu/smith/articles/lfo.html
* BFO: Basic Formal Ontology
http://www.ifomis.org/bfo/home
* Barry Smith, home page
http://ontology.buffalo.edu/smith/
* Barry Smith - Publications
http://ontology.buffalo.edu/smith/BSPUBLST.htm
[12] http://www.jfsowa.com/krbook/
[13] http://www.w3.org/TR/rdf-mt/
[14] http://en.wikipedia.org/wiki/Cyc
[15] http://www.opencyc.org/
[16] Wandering Significance: An Essay on Conceptual Behaviour
Mark Wilson, Oxford University Press 2006
http://www.oxfordscholarship.com/view/10.1093/acprof:oso/9780199269259.001.0001/acprof-9780199269259
[17] http://hnsttl.blogspot.com/2009/09/critical-notice-of-mark-wilsons.html
[18] http://www.topicmaps.org/ see link on page to:
[19] http://www.aw.com/samplechapter/0201749602.pdf
[20] www.ontopia.net/topicmaps/materials/tao.html
...........................................................................
Murray Altheim <murray10 at altheim dot com> = = ===
http://www.altheim.com/murray/ === ===
SGML Grease Monkey, Banjo Player, Wantanabe Zen Monk = = ===
Boundless wind and moon - the eye within eyes,
Inexhaustible heaven and earth - the light beyond light,
The willow dark, the flower bright - ten thousand houses,
Knock at any door - there's one who will respond.
-- The Blue Cliff Record
Alexander Johannesen
Murray,
Thanks, fantastic response, bookmarked and saved and would you please just friggin' start blogging again? We need you!
Love,
Alex