Re: HTML 5 and Distributed Extensibility

On 9/25/07, David Orchard <dorchard@bea.com> wrote:
> 3. Re-use of existing syntax and idioms for extensibility
> 4. A balance between simplicity of authoring and richness of functionality.

I hate to sound like a broken record on this topic, but both of these
are currently provided for by the GRDDL specification's creative reuse
of the profile attribute, which has been somewhat short-sightedly left
out of both the current HTML5 specification and the XHTML2
specification.

GRDDL is now a W3C specification [1] that works in XHTML 1.0, XHTML
1.1 and (for a slightly smaller subset of GRDDL processors - namely
those which pipe input through HTML Tidy or a similar conversion
utility) HTML 4.

For point three, GRDDL allows authors to specify a profile URI which
describes (in XHTML) machine-readable rules that provides extra
semantic information about element and attribute in a document. As for
simplicity of authoring, it is as simple as adding a namespace URI to
a document, albeit without the worry over prefixes.

For common microformats (hCard, hCalendar etc.), GRDDL will not be
particularly useful - but for scaling up the microformats approach of
embedding data in visible HTML, GRDDL will allow it to happen in a
predictable, author-specified way that scales up to the level where
individuals can define their own meta-data profiles.

I also completely support the extension of HTML through 'xmlns'-style
namespacing. The Microformats project has shown the power of
unintended extensibility of HTML. The HTML Working Group should do
everything possible to enable a variety of different approaches to
extensibility and embedded, machine-readable data.

[1] http://www.w3.org/TR/grddl/

-- 
Tom Morris
http://tommorris.org/

Received on Wednesday, 26 September 2007 10:48:42 UTC