W3C home > Mailing lists > Public > public-sws-ig@w3.org > November 2004

Re: OWL-S version 1.1 now available

From: Bijan Parsia <bparsia@isr.umd.edu>
Date: Sun, 21 Nov 2004 10:22:14 +0900
Message-Id: <C6DF0B28-3B5B-11D9-A708-0003936A0B26@isr.umd.edu>
Cc: public-sws-ig@w3.org
To: Ian Dickinson <ian.dickinson@hp.com>

On Nov 21, 2004, at 9:26 AM, Ian Dickinson wrote:

> Daniel Elenius wrote:
>>  [snip]
>> And also
>> ds:ClosedList a owl:Class
>>             ; rdfs:subClassOf ds:List .
>> right?
> 'Fraid not.  The first/rest/nil construct already gives you a closed 
> list.

Doesn't stop it from being a subclass, of course :)

Alas, first/rest/nil doesn't quite do the job without an "aware" 

> The  most common way to do a bag-like construct in OWL/RDF is to use a 
> repeated property, but this has the drawback (some say, others 
> disagree) that under the standard open-world assumption, you never 
> know how many values there are:

No. Just add a restriction (or use a oneOf).

> ex:foo
>     a ex:Bar
>     ; ex:hasArg "1"
>     ; ex:hasArg "2"
>     .
> How large is the bag of arguments of foo? Answer: no less than 2.

Well, in this case, only because of builtin he inequality of the 

> If, however, you encode the args as a list, it's a closed construct 
> and you know exactly the size of the list (or bag, if you mean to 
> interpret it that way).

ex:foo a Restriction onProperty ex:hasArg cardinality = 2.

Now you know :)

Alternatively (but nastier).

fooargs oneOf {:a :b}. :a differentFrom :b.
ex:foo a Restrictin nProperty ex:hasArg allvaluesfrm fooArgs.

That only gives you a maxcardinality, natch.

Personally, I strongly feel that using lists like that is way way way 
way worse style. It doesn't really give you want you want, and in most 
cases it's not necessary.

> If that's an important distinction to be made, then yes.  I'm simply 
> advocating that the distinction not be made solely by parsing the name 
> of the class.  URI's are supposed to be opaque :-)

It's not made by *parsing* the URI. We use two URIs to indicate two 
(possibly) distininct concepts and use simple URI comparison to 
determine which concept we're dealing with (although a vanilla owl 
reasoner can't determine that they are distinct).

Bijan Parsia.
Received on Sunday, 21 November 2004 01:22:25 UTC

This archive was generated by hypermail 2.3.1 : Tuesday, 6 January 2015 20:54:14 UTC