W3C home > Mailing lists > Public > www-rdf-rules@w3.org > February 2007

Re: AtomList infinite or cyclic in all models

From: Bijan Parsia <bparsia@isr.umd.edu>
Date: Thu, 15 Feb 2007 10:30:09 +0000
Message-Id: <9171ABC3-021C-4CCB-B06C-454C1DF535F5@isr.umd.edu>
Cc: Ian MacLarty <iml@missioncriticalit.com>, www-rdf-rules@w3.org
To: Dave Reynolds <der@hplb.hpl.hp.com>

On Feb 15, 2007, at 9:57 AM, Dave Reynolds wrote:

> Ian MacLarty wrote:
>> Hello,
>> In the owl file http://www.daml.org/rules/proposal/swrl.owl, the
>> AtomList class is a subclass of rdf:List and places cardinality
>> restrictions of 1 on rdf:first and rdf:rest.  Furthermore rdf:rest  
>> for
>> AtomList is constrained to the range AtomList.
>> This seems to imply that in all models, AtomList will be either  
>> infinite
>> or cyclic.
>> Could someone comment on this definition of AtomList?
>
> This is a list for the RIF rules working group

No, it isn't:
	<http://lists.w3.org/Archives/Public/www-rdf-rules/>

"""This mailing list is intented for the discussion of queries and  
rules for RDF data. We invite practical discussions with the goal of  
coordination and shared understanding of other implementations. See  
the list of lists for related mailing lists and a description of  
their intended content."""

and:
	<http://lists.w3.org/Archives/Public/www-rdf-rules/2001Jul/0000.html>

"""This is a subgroup of RDF Interest to discuss rules and query for  
RDF and
DAML+OIL.  It may later recommend that a working group be started for
Rules/Query.""""

> and not a support list for SWRL.

It is, of course, not a support list of any kind, but a discussion  
list. However, most W3C discussion lists that I've been on welcome  
*this* sort of question.

> However, you'll see that that file also includes the statement:
>
>    <swrl:AtomList rdf:about="&rdf;#nil"/>
>
> So the cardinality restriction on rdf:rest on AtomLists can be  
> satisfied by a nil so there is no problem with finite lists.

Though, I'll point out, that, being a first order characterization,  
you cannot rule out models with "weird bits" related to your list in  
all cases.

I'll also note that AtomList is intended to be part of the syntax of  
the rules. I generally think mixing that in with domain  
considerations is unwise.

Cheers,
Bijan.
Received on Thursday, 15 February 2007 10:30:10 UTC

This archive was generated by hypermail 2.4.0 : Friday, 17 January 2020 22:46:20 UTC