Re: on Use cases/ Web portals

Gary, please see my in-line responses....

Gary Mosher wrote:
> 
> Jeff,
> 
> [[[ I'm not sure I understand your point. Are you saying that in order
> to determine if something is a mammal, you must decide if it has hair,
> is warm-blooded, and gives live birth to its young? If so, the ability
> to state these kinds of definitions is the sort of thing we hope to provide
> with OWL.]]]
> 
> My point was that knowing how much of something, something is, can be very
> useful information. Simply describing me as tall, blond, and handsome would
> the more useful if facts regarding "how" tall and "how" blonde and "how"
> handsome was attached to these words .

Sorry, I misunderstood you. You can certainly create ontologies that
allow you to express these qualifications. In fact, you can do this even
with RDF. That is, you could create a property called height whoose
values are numbers. Users could then ask for people whose height
exceeded a certain value, or from the persepctive of an ontology, a
person could be considered tall if their height exceeded a certain
value. Although blond and handsome are more difficult to quantify, you
could assign your own numeric scales to these as well.

> [[[ However, your point about publications seems somewhat different. It
> seems to say that people would be more interested in publications that
> have higher circulation. ... I might wish to rank things in different orders
> than you.]]]
> 
> My point is less about ranking and more about "qualifying" descriptive words.
> I just used circulation as a convenient example, but something like the age
> of the publication might be more useful information to attach to the descriptive
> word. It should also be understood that the "system" would allow the individual
> making the request to decide how important "high relevancy" on a key word should
> be. In other words, the ability to request only articles published in newest
> 30% of publications would be made possible by simply adding a number qualifier
> to the key word request.  For example: articles, publications-3, etc.

One of the key ideas of RDF and WebOnt is to go beyond simple keyword
search. We want to enable search that uses structured queries (like you
might use to access a database). Although we are not defining a standard
query language for WebOnt at this point, a query of the type you
mentioned would be feasible with WebOnt, assuming you knew the year of
publication for each article. With the right query tool or query
language, you could even do this with DAML+OIL (see www.daml.org), which
is a predecessor of WebOnt.

> [[[ Since this is really a mailing list for specific comments on the WebOnt
> requirements document, we are not likely to get much discussion on this topic
> here. I recommend that you post your idea to www-rdf-logic@w3.org.]]]
> 
> Being frank: If how you have mapped out these discussion boards is an example
> of what you intend to do for web navigation--I'm not feeling very optimistic.
> 
> Gary Mosher

Received on Monday, 25 March 2002 14:57:57 UTC