RE: [RIF] ACTION-944: Summarize issue of imports argument and propose a solution

> Do you imagine a Link would ever make sense to have somewhere inside a
> Condition, in some RIF dialect?  If so, what would it mean?

Yes, a locator IRI could be used in a RIF dialect with 'Implicational
Goals'.
A version of these is described, e.g., in Joshua S. Hodas' and Dale
Miller's
Representing Objects in a Logic Programming Language with Scoping
Constructs:
"When an implication is to be proved, the term on the left of the
implication
is added to the current program, and an attempt is made to prove the
term on
the right."
(http://citeseerx.ist.psu.edu/viewdoc/summary?doi=10.1.1.3.423)

The antecedent of an implication inside a Condition could either be
an explicit Group or an IRI of type rif:link dereferencing to a Group.


>                                                              How would
> it be different from a Const?

A Link -- or Const of type rif:link, <Const type="&rif;link">IRI</Const>
--
differs from a Const of type rif:iri, <Const
type="&rif;iri">IRI</Const>,
in that it is interpreted as the Group to which its IRI dereferences.
The notion of IRI dereferencing can refer to the W3C Recommendation
Architecture of the World Wide Web, Volume One:
"3.1. Using a URI to Access a Resource"
(http://www.w3.org/TR/webarch/#dereference-uri)


This (Const of type rif:link) interpretation
can also be used for <Import>.
Within the <location> and <profile> role tags,
we initially had <Const> tags of type rif:iri:

<Document>
  <directive>
    <Import>
       <location>
         <Const type="&rif;iri">IRI</Const>
       </location>
       <profile>
         <Const type="&rif;iri">IRI</Const>
       </profile>?
    </Import>
  </directive>*
  <payload>Group</payload>?
</Document>

I would now rephrase things so that the <Const> here is
not of type rif:iri but of type rif:link, as follows:

<Document>
  <directive>
    <Import>
       <location>
         <Const type="&rif;link">IRI</Const>
       </location>
       <profile>
         <Const type="&rif;link">IRI</Const>
       </profile>?
    </Import>
  </directive>*
  <payload>Group</payload>?
</Document>

We would thus have rif:iri, rif:link, rif:local,
and xs:... constants.

Adding the value "&rif;link" to the attribute type
of the element <Const> is even easier in BLD and PRD
than introducing a new element <Link>.

Harold


-----Original Message-----
From: Chris Welty [mailto:cawelty@gmail.com] 
Sent: December 10, 2009 12:53 PM
To: Boley, Harold
Cc: public-rif-wg@w3.org
Subject: Re: [RIF] ACTION-944: Summarize issue of imports argument and
propose a solution


Harold,

Could you take a look at this?  We need to come to closure on this
issue.

-Chris

Sandro Hawke wrote:
>> ACTION-944 is about the XML syntax of the <Import> type tag
>> and possible related classes that RIF dialects may require.
>> The <Import> class contains a <location> role tag and an
>> optional <profile> role tag.
>>
>> Within those <location> and <profile> role tags, we initially
>> had <Const> type tags:
>>
>> <Document>
>>   <directive>
>>     <Import>
>>        <location>
>>          <Const type=3D"&rif;iri">IRI</Const>
>>        </location>
>>        <profile>
>>          <Const type=3D"&rif;iri">IRI</Const>
>>        </profile>?
>>     </Import>
>>   </directive>*
>>   <payload>Group</payload>?
>> </Document>
>>
>> We then noticed that the IRI content here is not a rif:iri
>> but an xs:anyURI. So this calls for a class different from
>> <Const>, which could be called <Link>, as follows:
>>
>> <Document>
>>   <directive>
>>     <Import>
>>        <location>
>>          <Link type=3D"&xs;anyURI">IRI</Link>
>>        </location>
>>        <profile>
>>          <Link type=3D"&xs;anyURI">IRI</Link>
>>        </profile>?
>>     </Import>
>>   </directive>*
>>   <payload>Group</payload>?
>> </Document>
> 
> Do you imagine a Link would ever make sense to have somewhere inside a
> Condition, in some RIF dialect?  If so, what would it mean?  How would
> it be different from a Const?
> 
>      -- Sandro
> 
> 

-- 
Dr. Christopher A. Welty                    IBM Watson Research Center
+1.914.784.7055                             19 Skyline Dr.
cawelty@gmail.com                           Hawthorne, NY 10532
http://www.research.ibm.com/people/w/welty

Received on Friday, 18 December 2009 23:12:54 UTC