Leo wrote:
________________________________________
From: ontolog-forum-bounces@xxxxxxxxxxxxxxxx
[ontolog-forum-bounces@xxxxxxxxxxxxxxxx] On Behalf Of Obrst, Leo J.
[lobrst@xxxxxxxxx]
Sent: Wednesday, November 28, 2012 4:33 PM
To: [ontolog-forum]
Subject: Re: [ontolog-forum] doing standards [was - Re: Webby objects] (01)
Actually, if you look at SKOS, it was intended to represent (structured)
vocabularies which could then easily be mapped to ontologies. (02)
[EJB] Easily mapped to very weak ontologies. The SKOS notion of
'broader/narrower concept' isn't necessarily ontological subsumption. (03)
The simple way people use SKOS is to import SKOS and then use skos:altLabel for
synonyms for the ontology concept name represented in OWL/RDF. (04)
[EJB] Yes. This is a useful practice. (05)
A gensymed name for an ontology class is simply a URI without any human words!
I suggest you have a canonical name (it’s just a label) for an ontology
concept, but use vocabularies to map to that name, because in fact ontologists
(at least most of them so far) are humans using human language, and “x23892”
is not as good as “Person”. (06)
The non-ontologists will be confused, of course, but they currently confuse
terms with concepts, (07)
[EJB] To quote from SBVR, there are two 'reference schemes' (means of
identifying individuals) for the concept 'concept': a term and a definition.
You have to associate one or the other or both with the magic URI when you
create it; otherwise it has no semantics. The SBVR copout for providing only a
term is to say: When no definition is provided, the common understanding, as
evidenced by the NODE definition, for example, is intended. The problem here
is that there are primitive concepts in any ontology, and there is reliance on
the term and "common understanding", intuition, or on a definition that is not
definitive: a set of necessary and sufficient conditions. (08)
and think that they don’t need semantics because their column name or XML
Schema entity is labeled Person, and of course everyone knows what that means. (09)
[EJB] Well, the audience would probably be hard-pressed to understand a formal
definition of 'person', in that the definition will introduce a larger set of
terms for which they have no intuition. (010)
I.e., they don’t understand semantics, let alone ontology. Constructs in
structural models such as relational databases, XML schemas, UML models are
vocabularies. (011)
[EJB] I fully agree. But I have to say that the same is true of most OWL
models. Many, if not most, of the concepts are 'primitive', and the set of
necessary properties is rarely sufficient to characterize the thing, or if it
is, then most of those properties are primitive. It seems to me that all of
our formal power notwithstanding, domain ontologies are not like mathematics.
There is not a small fundamental set of primitive concepts from which a whole
theory can be axiomatically constructed. A typical domain ontology has a few
hundred concepts, and about half of them are at most partly characterized
axiomatically. All the rest appeals to common understanding of the term used
for the concept, or to the common understanding of the terms used in the NL
definitions ("annotations"). (012)
[EJB] Further, I would argue that common upper ontologies are useful only as
contributing partial characterizations of classes, and almost completely
useless in characterizing properties. What you do get by subtyping an upper
ontology concept is a set of axioms that do apply to the concept at hand -- a
set of necessary characteristics that are not sufficient for anything but the
high-level abstracton -- and will be useful to your reasoner. (013)
In addition, it gets even more complicated, i.e., the vocabulary/ontology
issue, in that many times other vocabularies need to be captured, i.e., a
purchaser of some type (a technical purchaser) refers to a product as X,
whereas a purchaser of another type (a non-technical purchasing agent) refers
to it as Y, and a customer may similarly have different types (technical,
non-technical), say for engineering products. So really what you need is a
vocabulary, a context of use, and the ontology. (014)
[EJB] There is the famous tale of the chemical plant in which there were two
colors of coverall: the yellow ones were for everyday protection, the blue
ones were vapor-proof, to be used in case of hazardous leaks and spills. Of
course, the operators all knew the importance of the color, and of course they
referred to the vapor-proof coveralls as "blue suits". There came a day when a
new purchasing agent was directed to acquire a new batch of "blue suits" and
acquired ordinary coveralls that were dyed blue. 4 operators were burned but
no one was killed. (015)
At VerticalNet (business-to-business e-commerce) in 2000 we had this issue, but
never resolved it before we went out of business. Domain sales directors would
see ontology class labels and say “that’s not right, we don’t use that term”,
mistaking vocabulary for concept. EVERY application indeed uses (or could) its
own vocabulary, depending on the nature of the application and its interface.
So we distinguished “presentation” from “representation”. Presentation varies
radically, but representation less so (you do have to capture the filagrees of
meaning in your ontologies, but how you refer to those “nodes” is relatively
independent). (016)
[EJB] Yes! We have, however, run into a serious problem in this area, in which
the standard XML tag cannot be interpreted by even senior software engineers if
it isn't the term used by his organization. One of the OAG specifications has,
in the annotations in the schema, a list of synonymous terms used in various
companies in the industry, hoping that one of them will trigger the correct
response. But often it triggers the reaction that these terms identify
multiple somewhat different concepts ("filagrees") to the same readers. This
problem, however, filters into the literature of the field as well, which means
it is not restricted to knowledge engineering and data exchange. (017)
[EJB] As Eliot Kinder, one of the Hypertext gurus, once put it: "It all comes
down to people talking to people, no matter how much cool technology we throw
at it." (018)
Thanks,
Leo (019)
--
Edward J. Barkmeyer Email: edbark@xxxxxxxx
National Institute of Standards & Technology
Engineering Laboratory -- Systems Integration Division
100 Bureau Drive, Stop 8263 Office: +1 301-975-3528
Gaithersburg, MD 20899-8263 Mobile: +1 240-672-5800 (020)
From: ontolog-forum-bounces@xxxxxxxxxxxxxxxx
[mailto:ontolog-forum-bounces@xxxxxxxxxxxxxxxx] On Behalf Of Andries van Renssen
Sent: Wednesday, November 28, 2012 8:30 AM
To: '[ontolog-forum] '
Subject: Re: [ontolog-forum] doing standards [was - Re: Webby objects] (021)
Leo, Doug, Amanda,
Separating NL vocabularies from Ontology terms is one option, but IMHO not the
best one.
Creating a separate Ontologies language that only maps to NL terms
unnecessarily isolates the ontology world from the application world.
There are solutions to solve synonym and homonym management and use without
such a separation. (022)
Andries (023)
On Tue, Nov 27, 2012 at 5:22 PM, doug foxvog
<doug@xxxxxxxxxx<mailto:doug@xxxxxxxxxx>> wrote:
On Thu, November 22, 2012 16:59, Obrst, Leo J. wrote:
> Sure, Amanda, and that's why I (and we) advocate using natural
> language vocabularies that are linked/mapped to ontologies.
If you are referring to advocating the separation of NL vocabularies from
ontology term names, we certainly agree. The ontology needs to express
an N-N mapping between NL terms and ontology terms. (024)
> This was a hard lesson
> learned (initially, by others, before my time) in the DoD in the early
> 1990s, and that I personally experienced in the metadata wars of the
> 2000s, where people will fight to the death to include their "words",
> mistaking these for the concepts behind them.* (025)
_________________________________________________________________
Message Archives: http://ontolog.cim3.net/forum/ontolog-forum/
Config Subscr: http://ontolog.cim3.net/mailman/listinfo/ontolog-forum/
Unsubscribe: mailto:ontolog-forum-leave@xxxxxxxxxxxxxxxx
Shared Files: http://ontolog.cim3.net/file/
Community Wiki: http://ontolog.cim3.net/wiki/
To join: http://ontolog.cim3.net/cgi-bin/wiki.pl?WikiHomePage#nid1J (026)
|