Bill, I wholeheartedly agree. Utility rather than dogma is a much
better criterion for evaluation.
At 04:55 25/01/2007, William Bug wrote:
Many thanks, Xiaoshu.
It's very helpful to get a sense of the full spectrum of opinion on
this issue.
I would agree for most all the folks on this list - myself included
- the most important aspect of an ontology is to provide a shared
semantics within a computational framework.
I don't believe, however:
1) That means the same thing to all "ontological engineers" - i.e.,
I think applications vary widely in how they construct and use on
ontology in a computational framework - e.g., Robert's earlier
statement, ",One can make an ontology in a formal language like owl,
but still be informal in the ontological distinctions made"
2) I don't believe that is ALL an ontology is.
In reference to David Booth's earlier comment re: the redundancy of
"formal" ontology - it was wonderful to hear someone else say that,
for I've often felt my intended use for an ontology (and the
requirements that engenders) DOES in fact make "formal" a redundant
adjective. The problem comes with point '2' above in this sense -
what ontology implies to me may need to be explicitly stated for
those to whom ontology does not carry that intrinsic property. As
Robert stated most succinctly, not all ontologies are expressed
using a mathematical formalism even when they are ontologically
formal - and visa versa
The Google results returned by "define: ontology" are equally
illuminating - and frightening. The authors of these pages are
truly braver and more knowledgeable souls than I - which implies -
though the pronouncements I make regarding the development and
intended use of ontologies MAY be necessary they are in no way
sufficient to define the class "ontology"
In the end, whether an artifact designed to promote a shared
semantics IS an ontology is less important than whether it can truly
support achieving the goals to which you apply it, whether you are a
philosopher, biomedical informaticist, or a car mechanic.
Cheers,
Bill
On Jan 24, 2007, at 11:09 PM, Xiaoshu Wang wrote:
Well, I think the discussion is good, but trying to define
"exactly" what an ontology is will always be a futile
attempt. Just like any concept, we all actually know what we are
talking about but cannot give it a precise definition.
Nevertheless, does it really matter if we can define what an ontology is?
For me, an ontology is just an engineer artifact created to be
shared. If an ontology cannot be shared engineeringly, it is
useless. For instance, can we consider an "ontology" defined in
OBO to be an "ontology" in the semantic web? I think not because
if so, how an RDF engine understand it. So pragmatically in an RDF
world, anything in RDF is an ontology because it does not matter if
it is an "ontology" or a "dataset", an RDF engine would have
treated them in the same way. Consider the following two statement
about "<http://example.x>http://example.x";.
1.
http://example.com/x rdfs:subClassOf
<http://example.com/y>http://example.com/y 2.
http://example.com/x a http://example.com/c
Will there be any different treatment for an RDF engine? They have
to dereference the same URI and reason them accordingly,
right? Does it matter if we label one as an "ontology" and the
other "not"? This is the reason that I still cannot understand the
motive behind the design of an owl:Ontology, it serves no purpose
whatsoever. Cheers
Xiaoshu
William Bug wrote:
That's much better for Wikipedia than getting too deep into ABox and TBox.
Thanks, Kei.
On the other hand, some may not agree with the focus on the
lexicon - "Ontology is defined as a formal specification of a
vocabulary, including axioms relating the terms" - though I do
like the accessibility of that description.
Of course, you could additionally reference the Wikipedia entries
for Abox & Tbox:
<http://en.wikipedia.org/wiki/ABox>http://en.wikipedia.org/wiki/ABox
http://en.wikipedia.org/wiki/TBox
Cheers,
Bill
On Jan 24, 2007, at 10:40 AM, Kei Cheung wrote:
Just to add to Bill's comments. According to the following paper:
<http://www.springerlink.com/content/hnn72w7r18238467/>http://www.springerlink.com/content/hnn72w7r18238467/
Ontology is defined as a formal specification of a vocabulary,
including axioms relating the terms. A dataset is defined as a
set of facts expressed using a particular ontology.
-Kei
William Bug wrote:
I think you are right, David - axioms would be better, as
algorithms implies - though doesn't proscribe - an
implementation strategy that may not be relevant to all uses of
formal ontologies. Perhaps the use of algorithms relates to Tom
Gruber's oft quoted description of what an ontology is - a
description that does not fit for everyone using formal ontologies.
Maybe some mention of how formal ontologies are used to test
formal assertions and some mention of the difference between the
TBox & the ABox (using more accessible expressions) would be useful as well.
Again - thanks for trying to put this out there. I do think it
can be a very useful resource.
Cheers,
Bill
On Jan 24, 2007, at 10:03 AM, David Decraene wrote:
I'd like to comment on these statements:
Perhaps it can be phrased better, but 'algorhythms' refers to
the fact that a formal upper level ontology has built-in
DISJOINT (and other) axioms which reflect back onto their
children (ergo the consistency check phrase). Axioms is perhaps
a better choice.
Also, the formal in formal ontology has nothing to do with
the language of representation (perhaps that part can be
phrased better as well to avoid confusion) but to the
formalism (formality of the ontology as you refer to it) that
is embedded in the framework.
I do not disagree that this page can be improved further
(which is the purpose and strongpoint of wikipedia), but
explaining in laymans terms what a formal ontology is about is a challenge.
-----Original Message-----
*From:* [EMAIL PROTECTED]
<<mailto:[EMAIL PROTECTED]>mailto:[EMAIL PROTECTED]>
[mailto:[EMAIL PROTECTED] Behalf Of
*Robert Stevens
*Sent:* woensdag 24 januari 2007 15:45
*To:* Phillip Lord; Alan Ruttenberg
*Cc:* public-semweb-lifesci hcls
*Subject:* Re: [biont] Nice wikipedia page on ontology
/'d be inclined to agree with Phil. I don't where the bit about
"algorithms" has come from. The other mistake, I think, is not to
make the distinction between formality of language for
representaiton and the formality of the ontology itself. The
latter is, I think, a matter of the distinctions made. One can
make an ontology in a formal language like owl, but still be
informal in the ontological distinctions made.
/Formal ontological distinctions can be encapsulated in an upper
level, but upper level otnoogies are not necessarily
formal.... the phrase also explicitely refers to
upper level ontologies that
are formal in nature... Anyway, it is bad at almost any level
Robert.
,At 13:55 24/01/2007, Phillip Lord wrote:
>>>>> "Alan" == Alan Ruttenberg <[EMAIL PROTECTED]
<<mailto:[EMAIL PROTECTED]>mailto:[EMAIL PROTECTED]>
<<mailto:[EMAIL PROTECTED]>mailto:[EMAIL PROTECTED]>> writes:
Alan> Start at
<http://en.wikipedia.org/wiki/Formal_Ontology>http://en.wikipedia.org/wiki/Formal_Ontology
Alan> -Alan
Well, it starts of with this....
"A Formal ontology is an ontology modeled by algorithms. Formal
ontologies are founded upon a specific Formal Upper Level Ontology,
which provides consistency checks for the entire ontology and, if
applied properly, allows the modeler to avoid possibly erroneous
ontological assumptions encountered in modeling large-scale
ontologies. "
Almost none of which I would agree with.
Bill Bug
Senior Research Analyst/Ontological Engineer
Laboratory for Bioimaging & Anatomical Informatics
www.neuroterrain.org
Department of Neurobiology & Anatomy
Drexel University College of Medicine
2900 Queen Lane
Philadelphia, PA 19129
215 991 8430 (ph)
610 457 0443 (mobile)
215 843 9367 (fax)
Please Note: I now have a new email - [EMAIL PROTECTED]
<<mailto:[EMAIL PROTECTED]>mailto:[EMAIL PROTECTED]>
Bill Bug
Senior Research Analyst/Ontological Engineer
Laboratory for Bioimaging & Anatomical Informatics
www.neuroterrain.org
Department of Neurobiology & Anatomy
Drexel University College of Medicine
2900 Queen Lane
Philadelphia, PA 19129
215 991 8430 (ph)
610 457 0443 (mobile)
215 843 9367 (fax)
Please Note: I now have a new email - [EMAIL PROTECTED]
<<mailto:[EMAIL PROTECTED]>mailto:[EMAIL PROTECTED]>
Bill Bug
Senior Research Analyst/Ontological Engineer
Laboratory for Bioimaging & Anatomical Informatics
www.neuroterrain.org
Department of Neurobiology & Anatomy
Drexel University College of Medicine
2900 Queen Lane
Philadelphia, PA 19129
215 991 8430 (ph)
610 457 0443 (mobile)
215 843 9367 (fax)
Please Note: I now have a new email -
<mailto:[EMAIL PROTECTED]>[EMAIL PROTECTED]
