Re: worries about useMentionOp and how queries relate to rules and proofs

>I was a good boy and tried out the XSLT approach. Praise and cash
>donations cheerfully accepted.

Sheesh, I'm impressed. By you, that is, not by XSLT. D3 >> B2, I'd say.

More I think about this, more I like the idea of allowing these 
thingies, but putting them in a the query in a place where they 
really are forced to be on a special stage, as it were. How about 
something like an explicit 'filter' construct, so we might have

>SELECT ?annot ?author
>	 WHERE { (?annot dc:creator ?author)
>		 (?annot dc:created ?when) }
>	   FILTER ( isURI(?author) ||
		 ?when < xs:dateTime(20050101T00:00:00Z) )

The point being that 'AND' reads altogether to much like a 
conjunction, as though you were just tacking an extra piece on to the 
RDF pattern in the WHERE; whereas this lets us have warnings like 
"filters cannot be represented in RDF"

Dan & Jos, would this make you any happier?

Pat

>
>On Sat, Feb 05, 2005 at 12:42:59PM -0600, Pat Hayes wrote:
>>  >Anyhow, we do many such jobs to prepare and extract triples
>>  >from almost everything (even books written in PDF recently)
>>  >and also to consume them, to put them in SVG on pocket PC
>>  >etc, but for such jobs we simply use XSLT (and of course
>>  >Python, Java, ...)
>>
>>  Hmm, point taken. Well, then maybe what SPARQL should do is to
>>  explicitly allow XSLT constructions as part of query, to describe
>>  syntactic filters (? Does that make sense? Im a little out of my
>>  depth here.)
>
>It is feasible, but I think the practicalities make an argument for
>keeping the constraint clause in the SPARQL language.
>
>We could get rid of the constraints altogether and people could do the
>same work in XSLT. For instance,
>
>	SELECT ?annot ?author
>	 WHERE { (?annot dc:creator ?author)
>		 (?annot dc:created ?when) }
>	   AND ( isURI(?author) ||
>		 ?when < xs:dateTime(20050101T00:00:00Z) )
>could be
>	SELECT ?annot ?author ?when
>	 WHERE { (?annot dc:creator ?author)
>		 (?annot dc:created ?when) }
>+
><xsl:stylesheet version = '1.0'
>      xmlns:xsl='http://www.w3.org/1999/XSL/Transform'
>      xmlns:r="http://www.w3.org/2001/sw/DataAccess/rf1/result">
>
>   <xsl:template match="r:sparql">
>     <r:sparql>
>       <xsl:apply-templates />
>     </r:sparql>
>   </xsl:template>
>
>   <xsl:template match="r:head">
>     <xsl:copy-of select='.' />
>   </xsl:template>
>
>   <xsl:template match="r:results">
>     <r:results>
>       <xsl:apply-templates />
>     </r:results>
>   </xsl:template>
>
>   <xsl:template match="r:result[r:author/@uri
>	or translate(r:when/text(),'TZ:-','')
>	   &lt; translate('20050101T00:00:00Z','TZ:-','')]">
>     <r:result>
>       <xsl:apply-templates />
>     </r:result>
>   </xsl:template>
>
>   <xsl:template match="r:annot">
>     <xsl:copy-of select='.' />
>   </xsl:template>
>
>   <xsl:template match="r:author">
>     <xsl:copy-of select='.' />
>   </xsl:template>
>
>   <xsl:template match="*" />
>
></xsl:stylesheet>
>
>The constraint:
>	   AND ( isURI(?author) ||
>		 ?when < xs:dateTime(20050101T00:00:00Z) )
>turns into:
>   r:result[r:author/@uri
>	   or translate(r:when/text(),'TZ:-','')
>	      &lt; translate('20050101T00:00:00Z','TZ:-','')]">
>
>and all the variables in the final projection (annot, author) get
>enumerated. The rest of the template just keeps the result form
>intact.
>
>Benefits:
>   B1. re-use tools (XSLT).
>   B2. eliminate burden of SPARQL processors to do some fairly non-RDF stuff.
>
>Drawbacks:
>   D1. have to serialize as XML (no direct API calls get filtering)
>   D2. only works for XML serialization of result sets (doesn't filter
>       CONSTRUCT)
>   D3. need to learn an extra language (and I find XSLT difficult).
>--
>-eric
>
>office: +81.466.49.1170 W3C, Keio Research Institute at SFC,
>                         Shonan Fujisawa Campus, Keio University,
>                         5322 Endo, Fujisawa, Kanagawa 252-8520
>                         JAPAN
>         +1.617.258.5741 NE43-344, MIT, Cambridge, MA 02144 USA
>cell:   +81.90.6533.3882
>
>(eric@w3.org)
>Feel free to forward this message to any list for any purpose other than
>email address distribution.
>
>Content-Type: application/pgp-signature; name="signature.asc"
>Content-Description: Digital signature
>Content-Disposition: inline
>
>Attachment converted: betelguese2:signature 17.asc (    /    ) (00108F90)


-- 
---------------------------------------------------------------------
IHMC		(850)434 8903 or (650)494 3973   home
40 South Alcaniz St.	(850)202 4416   office
Pensacola			(850)202 4440   fax
FL 32502			(850)291 0667    cell
phayes@ihmc.us       http://www.ihmc.us/users/phayes

Received on Saturday, 5 February 2005 20:24:26 UTC