Pat, Matthew,
[The topic seems to have changed - would it be appropriate to change
the title?]
Saw my name mentioned, and would like to add a little to what Matthew
said.
I recall publishing this point in a couple of papers but cannot lay my
hands on them at the moment.
The idea starts from Lowe's definition of an ontology
as “the set of things whose existence is acknowledged by a particular
theory or system of thought.” (E. J. Lowe, The Oxford Companion to
Philosophy)
If one then digs into the history and philosophy of science (taking a Kuhnian view) then one can (roughly) identify a number of
characteristics of scientific revolutions theories:
·
Generality.
·
Simplicity.
·
Explanatory power.
·
Fruitfulness.
·
Objectivity.
·
Precision.
[It is useful to take this area as a starting point as it is a
well-researched area, and we can re-use its work.]
While historians and philosophers of science may argue about the
details of the characteristics (e.g. explanation), it seems to me there is some
kind of consensus that there is something to most of these terms. Of course,
there is also lots of scope for discussion.
The idea is that as the theory improves (undergoes a revolution) then the
set of things whose existence it acknowledges also changes – so these
characteristics of the theory are good indicators of the quality of the
ontology.
Taking the point made below, I think it is useful to give some idea of
purpose. My interest is in operational enterprise systems – so the
purpose of the ontology I would look at is to support automated processes
within the enterprise. In the absence of a better name for the feature these ‘scientific
revolutions’ improve, I have called it ‘sophistication’ –
and the characteristics the dimensions of sophistication. I am reasonably sure
that in this arena the six characteristics are useful indicators. I leave it to
others to investigate whether this is true in other areas.
PH> That is interesting; so a good ontology
provides
> insight as well as mere description? I
agree that
> makes sense. And one might
even approach
> measuring 'insight' by the number of useful
> extensions it has, or elaborations it
supports (?)
Taking scientific theories as examples, it is more that without
elaboration or extension, the theory supports areas that were not considered by
the originator.
In general, it is quite important to clarify what each of the characteristics
are.
My experience in working with business ontologies re-inforces the ideas
that these characteristics are useful indicators.
One example that clearly sticks in my mind was in a system that used an ontology for corporate changes.
A few weeks after implementation, in discussion with the users they
showed us how the system was being used to track quite complicated Swiss
equities. Our initial re-action was to tell the users off for not making clear
that it was a requirement. However, it soon became clear that the underlying
ontology allowed for these to be supported without workarounds.
> >Finally, as a tie breaker I would appeal to elegance and
simplicity.
> >Admitedly, these are hard to pin
down, but, given that, it
> is surprising
> >how often people agree about what is more
elegant/simple.
>
> Yes, I have noticed that. And it seems to be
> largely independent of whether or not one
thinks
> it is correct.
MW: True.
>
It seems that this is in line with the Kuhnian
view above – point two. So at least some of the points would seem to be
relatively uncontroversial.
Regards,
Chris
-----Original Message-----
From: ontolog-forum-bounces@xxxxxxxxxxxxxxxx
[mailto:ontolog-forum-bounces@xxxxxxxxxxxxxxxx] On Behalf Of matthew.west@xxxxxxxxx
Sent: 01 February 2007 11:59
To: phayes@xxxxxxx
Cc: ontolog-forum@xxxxxxxxxxxxxxxx
Subject: Re: [ontolog-forum] Visual Complexity
Dear Pat,
Some clarifications below.
Regards
Matthew
> -----Original Message-----
> From: Pat Hayes [mailto:phayes@xxxxxxx]
> Sent: 31 January 2007 19:23
> To: West, Matthew R SIPC-DFC/D21
> Cc: ontolog-forum@xxxxxxxxxxxxxxxx
> Subject: Re: [ontolog-forum] Visual Complexity
>
>
> >Dear Pat,
> >
> >> >Hi
Everyone,
> >> >
> >> >... I do
think, though, that some
> >> >measure
of correction of logical constructions is probably also
> >> necessary,
> >>
> >> Amen to that.
But it is very hard to see how this is to be done. I
> >> REALLY wish
there were a nontrivial and useful notion of how to
> >> measure
'correctness' of an ontology. It is not enough to
> just say,
> >> it is correct
if it "fits the facts" in some sense, since
> ontologies
> >> may be based
on very different, possibly mutually contradictory,
> >>
conceptualizations, and yet both fit the facts perfectly well.
> >>
> >> I have been
worrying about this for years but have never
> managed to
> >> get my
thinking to a point where it seemed useful to
> pursue an active
> >> research
direction. If anyone has any ideas about it, I'd love to
> >> engage in a
discussion.
> >
> >MW: Well I know I have views on this, so here goes.
>
> I hoped this would happen :-)
>
> >
> >First, there are the basic quality principles:
> >
> >A quality ontology is "fit for purpose".
>
> Great. Chris Menzel said something like this
> also. Now I want to know what a purpose is. Not a
> definition, but some entries in a list of 'uses
> for ontologies'.
>
> >This means an absence of
> >defects, which are what prevents an ontology from being fit
> for purpose.
>
> Hmmm. That seems circular. What am I missing?
MW: Nothing. Defects are just the other side of the same coin.
>
> >An ontology that supported more purposes would be better
> than one that
> >supported a subset of those purposes.
>
> OK. Can I translate 'support a purpose' into
> 'entails a set of sentences'? So that entailing
> more is supporting more? Im going to guess the
> answer is NO. So, what other kinds of purpose can
> it be used for? Or do you want to leave this
> open, and just take 'purpose' as a kind of token
> word for an open-ended set? (Which is fine, Im
> just asking.)
MW: Well this was in the sense that the ontology is useful
(or not) for some purpose, which might loosely be interpreted
as an application (class). Notice I am stepping away from
correctness here. I'm fairly happy that whilst you can sometimes
prove something is wrong, it is pretty much impossible to prove
something is correct. An absence of evidence that something is
incorrect (absence of defects) is the best we can generally do.
MW: Examples of applications might be:
Answering questions based on common sense/common knowledge
Supporting the design of engineering artifacts
Supporting medical diagnosis
Supporting epidemiology
Providing a foundation for more specific ontologies
>
> >Clearly at this level, as you point out, two ontologies can
> fit the facts
> >but be wildly different. As the next level of differentiation
I would
> >look at the empirical/mechanistic degree of the ontologies.
> This distinction
> >comes from mathematical modelling. Nearly any curve can be
> modelled by a
> >polynomial _expression_, but this does not give you an insight
into
> >what is going on. On the other hand a mechanistic model is
based on
> >equations that give insight into how it works. I think the
> same principle
> >applies to ontologies, and would be reflected in the degree
> and nature of
> >the structure and axioms it contained.
>
> That is interesting; so a good ontology provides
> insight as well as mere description? I agree that
> makes sense. And one might even approach
> measuring 'insight' by the number of useful
> extensions it has, or elaborations it supports (?)
MW: This is what Chris Partridge at least likes to call fruitfulness.
A good ontology will often have things in it that turn out to be
unnexpectedly useful, and apply in places and ways that were not part
of the original intent, because of the depth they have.
>
> What bothers me is that one man's insight can be
> another man's conceptual mistake :-) Barry thinks
> (what he calls) the SNAP/SPAN distinction is
> insightful, I think its a philosophical error. We
> aren't likely to agree on what counts as
> 'structure', I'm going to guess :-)
MW: Well I agree with you here ofcourse about the coneptual mistake.
But
I think you are more likely to get Barry under simplicity/elegance,
because he has to duplicate a lot of stuff, such as a person becuase he
has to have two objects, one the physical object, and the other the
lifetime activity, and presumably a relationship between them.
MW: Again, I think looking for unexpected fruitfulness is the measure.
>
> >
> >Finally, as a tie breaker I would appeal to elegance and
simplicity.
> >Admitedly, these are hard to pin down, but, given that, it
> is surprising
> >how often people agree about what is more elegant/simple.
>
> Yes, I have noticed that. And it seems to be
> largely independent of whether or not one thinks
> it is correct.
MW: True.
>
> Thanks for the input.
>
> Pat
>
> >
> >
> >Regards
> >
> >Matthew West
> >Reference Data Architecture and Standards Manager
> >Shell International Petroleum Company Limited
> >Shell Centre, London
SE1 7NA, United
Kingdom
> >
> >Tel: +44 20 7934 4490 Mobile:
+44 7796 336538
> >Email: matthew.west@xxxxxxxxx
> >http://www.shell.com
> >http://www.matthew-west.org.uk/
> >
> >
> >_________________________________________________________________
> >Message Archives:
http://ontolog.cim3.net/forum/ontolog-forum/
> >Subscribe/Config:
> 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 Post: mailto:ontolog-forum@xxxxxxxxxxxxxxxx
> >
>
>
> --
>
---------------------------------------------------------------------
> IHMC (850)434
8903 or (650)494 3973 home
> 40 South Alcaniz St. (850)202 4416 office
> Pensacola (850)202
4440 fax
> FL 32502 (850)291
0667 cell
> phayesAT-SIGNihmc.us
http://www.ihmc.us/users/phayes
>
>
>
_________________________________________________________________
Message Archives: http://ontolog.cim3.net/forum/ontolog-forum/
Subscribe/Config:
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 Post: mailto:ontolog-forum@xxxxxxxxxxxxxxxx