Re: GRDDL and transformations (take 2) [#issue-whichlangs]

On 8/29/06, Harry Halpin <hhalpin@ibiblio.org> wrote:
>
> Danny Ayers wrote:
> >  But what about
> > if there was a long doc and a timeout on the processor, so in certain
> > circumstances only a subset of the full graph was produced - the
> > function has another input, time. Partial fulfilment of the request
> > could well be useful, but strictly speaking is unacceptable.
> My guess would be then that this particular concern is out of scope and
> implementation dependent, as we can't guarantee people's processors
> working.

Out of scope for that case seems fine, I just hope there isn't some
other case that might turn around and bite.

> > My concerns may be off the mark, or just totally irrational, but
> > something makes me itch with the idea of the XSLT processor (or
> > Javascript interpreter, whatever) being implicit in the chain.
> However, as I put out earlier in response to Liam's concern - if you
> have say, a URI at the end of transform link that ends with the
> "rdf/xml" type, then there's no reason not to just "use it" instead of
> process it. So no processor in the chain in that regard.
>
>  Again - that use-case is sheer speculation - I still think we should
> stick to XSLT 1.0 for SHOULD. Unless anyone else likes the "rdf/xml" type.

Yep. Although I rather like the possibility of other types, they do
seem exceptional and SHOULD for XSLT 1.0 seems the best fit. (Hmm, is
the case of multiple transformations covered? I need to re-read).

In this same area, I'm not entirely comfortable with "The use of the
XSLT document() function to incorporate other data at transformation
time is an error", it seems very arbitrary, but unless anyone else
raises it I'll defer.

Cheers,
Danny.

-- 

http://dannyayers.com

Received on Tuesday, 29 August 2006 20:31:51 UTC