You say it very precisely:
usual way is to map terminologies into ontologies.
Some ages ago it was usual to consider the earth to be flat.
I propose to change this non-engineering approach in the so-called ontology field to a discipline where adequate quality control and repeatability are both
first class citizens. And of course community specific terminologies and even more are part of such an engineering approach.
But that requires that the homonym problem as previously identified by yourself, and recently came up again on this list, is solved. However such a useful
endeavour seems not to be popular in the high falutin world.
In a recent email on this list one can read:
OK – now on the the sarcastic side of things:
I will vigorously argue my position on the Ontolog Forum until we all agree that the cause the the different points of views was once again a
failure to clarify the meaning of a single term.
Also John made a pertinent statement:
Many so-called ontologies should more properly be called terminologies.
There are no classifications of knowledge domains that are sufficiently precise that the definitions could be specified in logic and be used in formal reasoning.
Chief Technical Officer
Tel: +31 (0)88-777 0 444
Mob: +31 (0)6-21 510 844
Fax: +31 (0)88-777 0 499
Van: ontolog-forum-bounces@xxxxxxxxxxxxxxxx [mailto:ontolog-forum-bounces@xxxxxxxxxxxxxxxx]
Namens Obrst, Leo J.
Verzonden: maandag 11 februari 2013 0:46
Onderwerp: Re: [ontolog-forum] Human knowledge domains ontology
The usual way is to map terminologies into ontologies. Terminologies can represent many different user communities, who wish to preserve their
terms (words, phrases) but yet map into concepts (ontologies) that represent an approximation to and representation of what those terms mean. Of course distinct natural languages have the issue of mapping their terms to those concepts too. In general, the
mapping is n terminologies to 1 ontology, but realistically things can get more complicated.
It was said in the email below:
to enable structured modeling of terminologies, controlled vocabularies, and thesauri, so to map these into ontologies.
I believe this mapping is a non-repeatable performance, and is not an engineering act, in the sense that an engineering act has proper quality
control and repeatability
From a complete ontology, in the sense of ISO TR9007, it is possible to provide a projection that delivers structured terminologies, controlled
vocabularies and thesauri. Hence repeatable.
Unfortunately the other way around is not repeatable but extremely popular.
I agree with William. You are probably looking instead for a terminological (non-ontology) “subject area classification” system, along the lines
of the Library of Congress Classification (LCC) system and the Library of Congress Subject Headings (LCSH), ODP’s DMOZ, the Dewey Decimal Classification (DDC) system, the Universal Decimal Classification system (UDC), Chaim Zins’ Pillars of Knowledge, or something
comparable. Searching on each of these will provide URIs. Sorry, don’t have them right now.
The W3C SKOS standard was developed to enable structured modeling of terminologies, controlled vocabularies, and thesauri, so to map these into
ontologies. Term/terminology vs. ontology is the usual distinction.
I do think that upper ontologies do not mostly organize knowledge by domains of the human endeavors intended to extend our knowledge, which is what your examples below show.
Rather, upper ontologies are concerned with indidviual concepts, like attitudes, things, and events. One of these individual concepts might be domain of endeavor. I think or hope that the definion of such a concept would not lead to its subtyping by different
domains, like astrophysics, as part of the upper ontology.
I think domains of human knowledge have been considered most extensively in library science, and in university course catalogs, and in encyclopedias. (including the Wikipedia)
But this is not a nice hierarchical structure -- thus university courses and even departments change regularly.
And, so Librarians turned to faceted analyis in the early 1970s.
There are journals and conferences etc about these kinds of classifications.
The Library of Congress has a new subject classification starting in 2012. The Dewey Decimal System is a little rustier.
This email is confidential and proprietary, intended for its addressees only.
It may not be distributed to non-addressees, nor its contents divulged,
without the permission of the sender.