W3C home > Mailing lists > Public > public-xmlhypermedia@w3.org > August 2012

RE: use cases

From: Rushforth, Peter <Peter.Rushforth@NRCan-RNCan.gc.ca>
Date: Sun, 19 Aug 2012 11:53:05 +0000
To: David Carlisle <davidc@nag.co.uk>, "public-xmlhypermedia@w3.org" <public-xmlhypermedia@w3.org>
Message-ID: <1CD55F04538DEA4F85F3ADF7745464AF1AE2FC42@S-BSC-MBX4.nrn.nrcan.gc.ca>

From: David Carlisle [davidc@nag.co.uk]
Sent: August 19, 2012 6:49 AM
To: public-xmlhypermedia@w3.org
Subject: Re: use cases

On 19/08/2012 11:27, Rushforth, Peter wrote:
> Liam, David(s),
> I was hoping for some comments on this idea, which seems especially
> topical given the preference for the MicroXML CG to drop all
> colon-names.
> In particular, the concept of a 'virtual' link on say the document
> node might be a data model issue.  Is this conceivable?  My
> perspective is you can pretty much declare the semantics you want
> when you define a media type.  But if it is theoretically
> incompatible with the data model, maybe that presents a problem :-).
> In short, for others who might not want to read Liam's full paper,
> by combining a new XML media type with generic, XML web-style links
> + automatic namespaces, one can get rid of colon names without
> getting rid of the benefits of namespaces.
> Thanks Peter ________________________________________ From:
> Rushforth, Peter [Peter.Rushforth@NRCan-RNCan.gc.ca] Sent: August 2,
> 2012 1:03 PM To: public-xmlhypermedia@w3.org Subject: use cases
> Hi Folks,
> I've added more beef to the use cases for automatic namespaces and
> "NeoXML".
> Please have a look and let me know if there are details that could
> be improved.  I seem to be working in a layered fashion, but we can
> always go back and amplify something if it does not support the
> layer above.  I hope this will help build the case for each and every
> vowel proposed.
> So first read automatic namespaces, then read NeoXML:
> http://www.w3.org/community/xmlhypermedia/wiki/Distributed_extensibility_/_automatic_namespaces
> Send comments back here.
> Thanks Peter

I must be missing something because I don't understand this at all.
If a system can be configured to understand

rel="ns" href="http://www.w3.org/XML/1998/namespace"

Then it could be configured to understand href attributes generally, so
presumably it could be similarly configured to understand any other
hypertext related attributes such as src, it wouldn't really need the
namespace mechanism would it?

Sorry I think I don't understand your point of confusion. 
The namespace mechanism allows the mapping of unprefixed elements and attributes to namespace URIs via a file mechanism vs the standard xmlns + prefix mechanism.   I guess my hope would
be that the recognition of that particular value in the href together with either the @rel or @type, or both, would allow a processor to make the assumption that href, src etc in the instance document are defined/mapped by the resource at that URI, unless otherwise mapped by a more local rel="ns" href="foo.com" scoping.

When *combined* with a message that is transmitted with the application/nxml media type, (and its virtual declaration of those three attributes on the document node), yes, a processor could be safely programmed to rely on the semantics of href, src, type etc.

>In your automatic namespace document you appear to be assigning
>namespaces to uprefixed attributes so this would mean the
>documents were not expressible in xml 1.0 + Namesapces. That isn't
>necessarily bad but it ought to be highlighted if it is true.

Why would they not be expressible in xml 1.0 + Namespaces?

Received on Sunday, 19 August 2012 11:53:35 UTC

This archive was generated by hypermail 2.3.1 : Tuesday, 6 January 2015 19:44:53 UTC