I agree nouns are crucial to model, as classes though, definitely not via names of properties. Take noun Sister (oh brother -- looks like a transitive relationship!) -- itself an instance of Class. In the first place, instances of the concept sister sisterhood sisterlyness or other, can only be represented as a subject or an object (that is, if we're on the same page of a subject-predicate-object modelled-world, if not, then, wow!)
Pat mentions COSMOS' numerous subproperties of isaPartOf -- I happen to believe the adoptability of an ontology is inversely proportional to the number of assertions it proposes, so I wonder why he sees them "all ... useful in their proper context". As I am proposing that no nouns be (in) predicates, then the number of properties drops precipitously, with proper community focus then placed on nouns' models.
Go further: "X sister Y". There's a massive difference between "X sister-to Y" and "X sister-of Y" and "X sister-by Y" and "X sister-after Y" and "X sister-before Y" and "X sister-with Y" and so on. Let's model these notions as easily and simply as possible for people to understand - in that I agree alot with Pat.
These intricate relations can be modelled with has as the predicate with an instance of Sister as the object. This Sister instance has instances of the properties to, of, by, after, before, with and so on. This Sister instance can be related in a triple to a subject via only the predicates has, had, will have and so on. (I've been thinking lately there's need to distinguish has-a versus has-the...)
I much dislike the idea of an is-a property because as other notes here said, each "type" that a thing is, is really just a facade, a context, a role, etc, that the thing has at one time or another, so let's model it that way. Start with just one verb (has) as the only proper predicate -- its subproperties identify the critical elements of time deonism negation etc applicable to any relationship, elements which definitely should be specified in queries.
Thanks - jmc
On 20.05.2013 08:21, Patrick Cassidy wrote:
The use of simple nouns such as "mother" or "sister" by themselves to
designate a relationship presents a serious problem in determining the
polarity: in "X sister Y", who is the sister - X or Y? Even if one
ontologist used the polarity consistently, it would still be a problem for
others trying to understand the notation. I always use the verbal
"isTheSisterOf" or "hasSister" for such relations. As I understand him,
John Sowa also suggests such usage. As a final blow, I have seen
sophisticated ontologies using the simple noun where the intuitive
polarities are in fact different for different relations. Good grief!!
Although an ontology when used by a computer will do all and only what the
axioms specify, the ontologies have to be understood by people in order to
be used properly. What I consider "really, really bad practice" is to use
any labels other than those most easily and rapidly understood by the human
who is looking at the ontology. Useful ontologies are big, and it is a high
crime to force people to waste time trying to figure out what the intended
meaning is.
There is an opposite consideration: if labels are easy to understand, users
may think they understand them when in fact they don't. Well, that is the
lesser problem, solved by ontology users taking the time to carefully read
the description of an element at least once. On subsequent usage, the
labels serve as mnemonics to refresh the memory of proper use, which needs
to be carefully understood on first encounter. But if the labels are
themselves deceptive or ambiguous, the likelihood of forgetting proper
usage, even with the mnemonics, increases.
The matter of general or ambiguous relations such as 'isaPartOf' requiring
more specific subrelations is a different issue. There are lots of 'part'
subrelations in COSMO, and in other ontologies I have seen. All are useful
in their proper context.
Pat
Patrick Cassidy
MICRA Inc.
cassidy@xxxxxxxxx908-561-3416
-----Original Message----- From: ontolog-forum-bounces@xxxxxxxxxxxxxxxx [mailto:ontolog-forum- bounces@xxxxxxxxxxxxxxxx] On Behalf Of John F Sowa Sent: Sunday, May 19, 2013 11:30 PM To: ontolog-forum@xxxxxxxxxxxxxxxx Subject: Re: [ontolog-forum] is-part-of: a really, really, bad practice? On 5/19/2013 5:17 PM, jmcclure@xxxxxxxxxxxxxxwrote:
I argue the industry badly needs consensus about the best practice for how attributes/relations are to be named.
My recommendation is to use the most common phrase in ordinary language.
in "old style" systems these names are nouns, perhaps qualified nouns; in "new style" systems these names are, uh, something other than a
noun. In ordinary English, it's common practice is to represent relations with nouns. The syntax of English and other languages allows verbs and adjectives to be *nominalized* in order to refer to the relations: "The Romans destroyed Carthage in 146 BC" => "The destruction of Carthage by the Romans in 146 BC" "The book is easy to read" => "The ease of reading the book" For words like part and family relations like mother, child, sibling, uncle, etc., there are no obvious verbs. It's more convenient to use noun + 'of'. In fact, English syntax makes it easy to switch 'of' to 'has' in order to form inverses: "X is the father of Y" => "Y has a father X" (or has X as father). "X is a part of Y" => "Y has a part X" The nominal form is easy to modify as needed: "X is a proper part of Y" => "Y has a proper part X" "X is an only child of Y and Z" => "Y and Z have an only child X" This seems like a good argument for using nouns to name relations. John _________________________________________________________________ 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
_________________________________________________________________
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
_________________________________________________________________
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