Adding XML Namespace Support

[inadarei]

From @mamund on March 16, 2014 18:0

I started this issue to cover the PR @dret started (which we can no longer easily merge).

See this thread in the list for background: https://groups.google.com/forum/#!msg/uber-hypermedia/tDhTK57SBqk/c7Jv3vTbjzAJ

• decide on an NS value for UBER
• update the spec to make using the UBER NS a MUST
• update all the XML examples to be compliant

Copied from original issue: mamund/media-types#31

[inadarei]

From @dret on March 16, 2014 22:16

also, it should be specified if/how non-uber namespace elements/attributes can be used and have to be processed. i think it would be excellent to have a clear "processing model" section in the spec, and this would be one of the things that should go in there.

[inadarei]

From @mamund on March 16, 2014 22:18

@dret:

would you be willing to write up the processing model section for UBER? should this be part of the spec doc (Normative) or a secondary doc (Informative)?

i'd like to avoid getting too many implementation details into the doc but want to provide enough to make implementors life easier (@smizell proly has some things to say on this, too).

[inadarei]

From @dret on March 16, 2014 22:26

i don't think i understand everything well enough to write a processing model section. but to me, any spec that has such a section is much more useful than one that doesn't. also, with uber going the RDF way of being multi-syntax, each syntax should come with it's own processing model, so that there's a well-defined how how to go from representation to model and back. without that, it's pretty likely that guessing will be involved for some aspects of the serialization/parsing process, or at least it will be harder to find all the relevant parts in the spec.

[inadarei]

From @mamund on March 16, 2014 22:31

by "syntax" you mean format right (XML & JSON)?

and i understand about not wanting to take this on alone. would you be able to help me (or others) out as we write this up?

[inadarei]

From @dret on March 16, 2014 22:34

sure, more than happy to review. and yes, by syntax i mean representation of the model. many models have one representation, and some choose to have more than one. it complicates matters on a variety of levels, but can help as a way to overcome entrenched syntax communities.

[inadarei]

From @mamund on March 16, 2014 22:47

agreed.

i think we can go for IANA (vnd) registration before sorting out the process model(s) but want to make sure to settle the NS thing before registration.

sound like a plan?

[inadarei]

From @dret on March 18, 2014 22:34

i guess that's correct. but it might help to have a fully spec'ed media type for registration.