Media-types and GRDDL

I do think it would be good to add a sentence or so about media-types to
the GRDDL spec - it seems to be a fairly common misunderstanding (as was
just displayed by people as bright as DanBri on the SemWeb interest
list-serv [1])
that people have that GRDDL limits them to XSLT 1.0 and has no mechanism
for anything else. Well, it *might* - but we only require XSLT 1.0 - so
let's be clear about that.

So, I suggest adding the following sentences to the GRDDL spec somewhere:

"GRDDL implementations SHOULD support XSLT 1.0 as a transformation
language, and MAY support other languages."

"Languages GRDDL implementations are encouraged to support include
languages like XSLT 2.0 and XQuery."

Now, for the more controversial media-type sentence:

"GRDDL can use media types of the transformation resource to detect the
what type of language the transformation uses. For example, if in a
XHTML document the link element has an href attribute of
"http://www.example.org/transformation" and this URI returns a document
with a media type of "application/xslt+xml," GRDDL is licensed to run
the transformation using XSLT."

Notice I'm intentionally a bit vague about XSLT 1.0 or XSLT 2.0 here. I
believe they share the same media-type. So, to give a single option
(since Xquery doesn't have a media-type I think) and to answer this
possible issue, here's another tw more controversial sentence:

"However, if the resource had a media type of "application/javascript,"
GRDDL could run, if the implementation supported it, a javascript
transformation."

And:

"Since many languages do not have media types and multiple versions may
share a single media type, GRDDL implementations can use various
implementation-dependent techniques to attempt to discover the language
or version of a language used by a transformation. However, the only
language transformation required to be supported by a GRDDL
implementation is  XSLT 1.0."

I realize that last sentence is repetitive but we should repeat
ourselves sometimes to make points absolutely clear.

-- 
		-harry

Harry Halpin,  University of Edinburgh 
http://www.ibiblio.org/hhalpin 6B522426

Received on Tuesday, 5 September 2006 01:35:21 UTC