What is the nature of HTML 4.0? was RE: Proposal for new RDDL naturesand purpose

[Jonathan Borden <jborden@mediaone.net>]

Eric van der Vlist wrote:
>
>
> I'd like to propose a new "purpose" for RDDL that would be something
> like "alternative representation" (I think that it's slightly different
> than a "normative resource") and new natures that would cover commonly
> used formats (xhtml, html, wml, svg, ... as well as RDF).

XHTML is easy to assign a nature to, HTML is not so obvious in my mind:

XHTML
nature: http://www.w3.org/1999/xhtml

(in this case the namespace of the root element, namely 'html' is a good URI
to use as the 'nature' of XHTML.)

HTML

should the nature be the URI of its specification?

-- http://www.w3.org/TR/html4/

or should the nature be the well-known URI of its content-type: text/html
-- http://www.isi.edu/in-notes/iana/assignments/media-types/text/html

or should the nature be something from the DOCTYPE e.g. strict.dtd?

-- http://www.w3.org/TR/html4/strict.dtd

normally I don't like to use a specification as a nature, but in this case
the specification directory is also the root directory of the DTDs...

>
> My first motivation is to allow to specify the location of a RDF
> document that would be equivalent to the RDDL (in case its publisher
> would like to provide it).

this is the motivation behind RDDL itself, to allow the specification of
resources that provide either alternate representations of or schemata that
describe a namespace as well as other resources such as code, stylesheets,
transforms etc.

>
> I have also noticed that using XHTML or HTML namespaces as the nature of
> resource seems to be a common practice and I think it should be
> documented.

the resources you refer to are specifications that are intended to be human
readable.

>
> There is also a decision to take for these types of document about using
> the URI of their mime types or their namespace URI.
>

In general when the root namespace URI is adequate to describe the nature of
a resource this is the prefered nature.

Yet this doesn't always work for example a RDDL document itself has a root
element namespace of http://www.w3.org/1999/xhtml but this doesn't alone
describe the nature of a RDDL document. For a RDDL document:
http://www.rddl.org/ best describes its nature (from this URI a user or
program can get many many resources with which to manipulate it).

When a document isn't XML (and hence can't have a root namespace URI) its
well-known media type URI may serve as the nature.

-Jonathan