Chris, (01)
you wrote: (02)
> The possible semantic varieties of 'is-a' in natural language are quite well
> known.
>
> It seems to me that you are dealing with the subsumption variety here (Man
> is an Animal - all men are animals).
>
> Isn't there another variety that needs also to be considered - the
> instantiation variety. Socrates is a man. (I wonder how the UML-ites will
> diagram this?) (03)
In UML2, the notation for an individual is a box with an underlined
label of the form <thing name>:<class name>, e.g., Socrates:man. (04)
But the point is exactly that: what reference concepts and reference
relations do we want to specify as our base language? Which of those
will we actually have graphical representations for? And will those
graphical representations be clearly distinguished? (05)
> While one can see subsumption as a relation between instances of terms
> (Every instance of A is an instance of B) - one can also see it as a
> relation between the terms. If one takes the second course, then these two
> varieties of is-a share the property of being a relation between the terms
> that are linked to rather than of instances of the term. For example, that
> the class of men is a sub-type of the class of animals. (06)
I think this is what Matthew was saying, and what I disagreed with. (07)
> This property is also shared by some of the other relations you mentioned -
> union, intersection, etc. (08)
Yes. That was Pat Hayes' point. (09)
> - though again these can be cashed out in terms of their instances. (010)
Umm, I suppose. My point was that you don't want to confuse the
notation for "An X can have a friend who is a Y" with the notation for
"No X is a Y". (011)
> You mentioned relations that are between instances/members of the relata -
> e.g. A -- is part of --> 1..1 B. Won't the notation also need to deal with a
> -- is a part --> b or (is-part-of a b) (where a is my hand and b my arm).
> Here again the relation is relating the terms rather than instances of the
> terms. (012)
The relation is "organ is-part-of organ"; the member is a
sentence/proposition: "my-hand is-part-of my-arm". We need to be able
to tell that 'organ' is a term for a classifier, whereas 'my hand' is a
term for an individual. We need to tell that the first is a vocabulary
entry and the second is a proposition. And we need to tell that
'is-part-of' in the second is a reference to the vocabulary entry that
is the first. The formal language of the ontology must make such
distinctions and make them clear. OTOH, it is not clear to me that a
non-normative graphical language needs to be able to represent
propositions of the "my-hand is-part-of my-arm" kind at all. It depends
on whether we need such diagrams as an aid to comprehending the ontology. (013)
At the risk of opening Pandora's box even wider, I will point out that a
"class whose members are classes" shows that the role "member" isn't
necessarily played by an "individual". (Or one can say that some
classes are individuals, or whatever your modeling language allows.) In
particular, a concept that will appear in the units of measure model is
"quantity kind", with the property that each member of 'quantity kind'
is a subtype of 'quantity' (but not vice versa). Examples: "length",
"duration", "mass". (014)
The idea is to take these concerns into account in doing language
selection. My particular position is that a non-normative diagramming
language will help comprehension of the ontology, but it need not cover
more than the basic features of the ontology, and what it does depict
should not be confusing. (015)
-Ed (016)
"A picture is worth 1000 words, but it takes 10,000 words to
recover from the 500 words that you didn't mean but people saw in
the picture."
-- John Day (017)
--
Edward J. Barkmeyer Email: edbark@xxxxxxxx
National Institute of Standards & Technology
Manufacturing Systems Integration Division
100 Bureau Drive, Stop 8263 Tel: +1 301-975-3528
Gaithersburg, MD 20899-8263 FAX: +1 301-975-4694 (018)
"The opinions expressed above do not reflect consensus of NIST,
and have not been reviewed by any Government authority." (019)
_________________________________________________________________
Message Archives: http://ontolog.cim3.net/forum/uom-ontology-std/
Subscribe: mailto:uom-ontology-std-join@xxxxxxxxxxxxxxxx
Config/Unsubscribe: http://ontolog.cim3.net/mailman/listinfo/uom-ontology-std/
Shared Files: http://ontolog.cim3.net/file/work/UoM/
Wiki: http://ontolog.cim3.net/cgi-bin/wiki.pl?UoM_Ontology_Standard (020)
|