Re: XBL media type?

On Tue, 22 Aug 2006, Mark Baker wrote:
> 
> The first is that it defines no media type for itself, and instead 
> prescribes that all XML documents with a root xbl element are XBL 
> documents.

Note that this is just a definition of a term used mostly for exposition, 
it doesn't actually affect the semantics of the language.


> I believe it would be better to mint an XBL specific media type and 
> change the definition of "XBL document" to say simply that all documents 
> served with this media type are XBL documents.  This would, for example, 
> prevent a literal result element XSLT stylesheet for an XBL document 
> being confused with an actual XBL document.

I don't really understand how such confusion could arise (or for that 
matter, why anyone would ever do an XSLT transform into an XBL document, 
but I concede that someone might find a reason to do that).

However, it doesn't matter how we define "XBL document", MIME type or no 
MIME type. The only way the that the term affects the processing model is 
in the defaulting of binding selection when interpreting URIs to bindings. 
If someone specifically invokes a binding, then they clearly intend the 
document to be processed as XBL; they wouldn't point to an XSLT document 
in this kind of situation (doing so would never do anything useful).

With the exception of this defaulting, the XBL processing model is defined 
in terms of the DOM, which is independent of MIME types.

Thus I don't think it is necessary for us to define a MIME type beyond the 
generic application/xml. XBL will be found in XML sent with all kinds of 
MIME types.


> Related to this, section 3.4 (Aug 21 draft) prescribes that;
> 
> "If the URI contains a fragment identifier, it must point to a specific 
> binding element within an XBL subtree in the specified document (e.g. by 
> matching the binding element's id attribute) [...]"
> 
> I think it's important to make it clear that it's the media type of the 
> retrieved or cached binding document that determines how the fragment id 
> is to be interpreted.  So I'd suggest a minor edit; that we add "if the 
> binding document is an XBL document" at the end of the example.  Plus 
> when we define the media type, we need to say (as part of the 
> registration template) that frag ids map to xbl:id (plus whatever else 
> we decide about xml:id).

Fixed. Let me know if the new text is ok.

Cheers,
-- 
Ian Hickson               U+1047E                )\._.,--....,'``.    fL
http://ln.hixie.ch/       U+263A                /,   _.. \   _\  ;`._ ,.
Things that are impossible just take longer.   `._.-(,_..'--(,_..'`-.;.'

Received on Tuesday, 22 August 2006 22:14:08 UTC