Re: SPARQL: BASE IRI resolution

On Wed, 2005-09-07 at 03:44 +0200, Bjoern Hoehrmann wrote:
> * Dan Connolly wrote:
> >> I am not comfortable with the "abbreviation mechanism", see e.g.
> >> http://lists.w3.org/Archives/Public/uri/2004Jan/0009.html for why.
> >
> >There I see Fielding answering some questions about URI references.
> >I don't see anything that should make one uncomfortable. Please
> >elaborate or be more specific?
> 
>   The effect of this change, when implemented, has no impact on the
>   protocol *except* when a person is deliberately abusing the base
>   URI by assigning it an unrelated URI for the purpose of creating
>   an artificial shorthand notation for external references.
> 
> If you have a SVG document at http://foo.example.org/ with
> 
>   <svg:g xml:base='http://bar.example.org/d.png'>
>     <svg:image xlink:href='a.png' ... /><!--bar.example.org/a.png-->
>     <svg:image xlink:href='b.png' ... /><!--bar.example.org/b.png-->
>     <svg:image xlink:href='c.png' ... /><!--bar.example.org/c.png-->
>     <svg:image xlink:href='d.png' ... /><!--same document-->
>   </svg:g>
> 
> the document would be in error as the same document reference is not
> allowed;

In what way is the same document reference not allowed? Is that
something special to SVG?  I don't see how it's relevant to SPARQL.

And if that's an error, then surely 
  <svg:image xlink:href='http://bar.example.org/d.png' ... />
is also an error; i.e. the possibility of the error is orthogonal
to the use of abbreviation mechanisms.

>  base resource identifiers therefore are not simply some kind
> of "abbreviation mechanism".
> 
> >No; it just re-states 5.1.4 "If none of the conditions described
> >above apply, then the base URI is defined by the context of
> >the application."
> 
> What about 5.1.2 and 5.1.3? It seems that this redundant part should
> either be removed or more clearly state that all of 5.1 must be con-
> sidered when resolving relative resource identifier references.

All of the whole RFC must be considered; it's cited normatively.
Aren't the implications of 5.1.2 and 5.1.3 clear enough with the text
as is? The SPARQL QL spec is just saying "if the BASE isn't
given explicitly, it's taken from somewhere else; somewhere that's
not constrained by this specification."

If it's not clear enough as is, we'd appreciate it if you would
suggest some wording that you might find acceptable.

-- 
Dan Connolly, W3C http://www.w3.org/People/Connolly/
D3C2 887B 0F92 6005 C541  0875 0F91 96DE 6E52 C29E

Received on Wednesday, 7 September 2005 02:43:56 UTC