W3C home > Mailing lists > Public > public-owl-wg@w3.org > November 2007

Re: abbreviated IRIs - ISSUE-14 (was Re: owl minutes)

From: Alan Ruttenberg <alanruttenberg@gmail.com>
Date: Sun, 18 Nov 2007 01:41:53 -0500
Message-Id: <048784E0-CB72-4259-9BF2-2C3FB7D0D6D1@gmail.com>
Cc: public-owl-wg@w3.org
To: Peter F. Patel-Schneider <pfps@research.bell-labs.com>

> I'm not sure what the motivation was for slightly expanding on  
> QNames in
> the abbreviated IRI syntax, but not going all the way to CURIES.  Does
> anyone know?

The weren't going to move towards CURIES at all, but for a last  
minute complaint by your truly. They agreed to modify to accommodate  
abbreviations like: pmid:123004002, i.e. leading numerals, but did  
not want to go further in order to reserve the possibility that other  
characters that CURIES might have permitted, specifically, I believe,  
"/" could be used for other purposes, such as "path" delimiting  
characters.

-Alan

On Nov 12, 2007, at 7:17 AM, Peter F. Patel-Schneider wrote:

>
> I wanted to wait until I had a chance to look at the SPARQL spec,
> http://www.w3.org/TR/rdf-sparql-query/ before approving the resolution
> of ISSUE-14.
>
> The relevant part of the SPARQL spec appears to be Section 4.1.1.  I
> read this normative section fully, and came up with the following
> conclusions:
>
> 1/ The syntax for IRI_REF does not match the syntax for IRI  
> references.
> 2/ The abbreviated syntax for IRIs does not match CURIES, nor anything
>    else that I can tell, and will not generate all IRIs.
>
> However, this section, although it is marked as normative, is not
> complete.  (I seem to remember a comment by me on this before -  
> very bad
> document design - LET'S NOT DO THIS!)  To get a complete answer one  
> has
> to go to Appendix A, where deficiency 1 is alleviated.  However, I
> believe that deficiency 2 not alleviated.
>
> I'm not sure what the motivation was for slightly expanding on  
> QNames in
> the abbreviated IRI syntax, but not going all the way to CURIES.  Does
> anyone know?
>
> I don't view it as an acceptable solution - if we are
> going to generalize QNames, then we should go all the way to CURIES.
>
> peter
>
Received on Sunday, 18 November 2007 06:42:05 GMT

This archive was generated by hypermail 2.2.0+W3C-0.50 : Tuesday, 8 January 2008 14:13:27 GMT