EBNF grammar question...

Sandro Hawke wrote:
> 
> Gary was kind enough to draft a response to WL's Last Call comment.  I
> added a header/footer.  Please look it over; I guess we'll send it
> Monday if no one has any comment.
> 
>     http://www.w3.org/2005/rules/wiki/Response_to_WL
> 
> (Chris, have you made any progress with the other LC comments?  We
> wanted to try to close the loop on each of them by Tuesday!)
> 
>      -- Sandro
> 

grammar question...

I just realized that our grammar has some maybe not so nice feature...
For 'Name's we allow arbitrary UNICODESTRINGs

Even if that doesn't seem to raise ambiguities per se, some things look 
awkward...

e.g.

1) in the Prefix production, it may look awkward, but doesn't 
necessarily create an ambiguity:

  Prefix( assd asdnewr asdswnke <http:example.org> )


nonetheless such prexif is not usable later on, since the CURIE 
production doesn't allow non-PN_PREFIXes [1] as prefixes


2) Moreover, it seems to me that this is a more severe problem for named 
arguments, e.g. according to the grammar, this one is valid, isn't it?

  _predicate( assd asdnewr asdswnke -> "Hello" )


Question/Suggestion: Can/shall we make the 'Name' production more 
restrictive?

Axel

	
1. http://www.w3.org/TR/rdf-sparql-query/#rPN_PREFIX
-- 
Dr. Axel Polleres
Digital Enterprise Research Institute, National University of Ireland, 
Galway
email: axel.polleres@deri.org  url: http://www.polleres.net/

Received on Tuesday, 11 August 2009 11:09:18 UTC