Re: C implementation of DOM

[Bill Haneman <bill haneman sun com>]

Michael Meeks wrote:
> 
> Hi Jason,
> 
> On Mon, 2002-04-01 at 01:24, Jason White wrote:
> > At the CSUN Linux access session, it was indicated that one obstacle
> > preventing Gnopernicus and similar applications from taking advantage
> > of the semantics offered by the W3C Document Object Model, was the
> > lack of a C-based API for the DOM.
> 
>         Is that so ?

Well, we believe the lack of a (standard) C binding for DOM to be a 
significant issue... not so much for gnopernicus as for accessibility
of web content.

However there are a host of issues relating to view-versus-content
accessibility here, and I'd sum up to say that DOM-based access (to the
actual content DOM or something closely based on it) is a desired
adjunct to access to the view "DOM".  ATK and AT-SPI provide a nice
API for the "view object" (we hope ;-) but cannot expose a
w3c-style DOM for the document object readily.  In any case the 
"view" object is central to most accessibility use cases.

> > I was reminded of this issue upon
> > reading a freshmeat.net announcement of an implementation of the DOM
> > in C: http://users.erols.com/mballen/domc/
> >
> > This claims to implement DOM levels 1 and 2 core, as well as DOM level
> > 2 events.
> 
>         As I understand it there are several projects that do DOM like things
> the details of which I'm ignorant of, but Daniel will know - Daniel ?

There were gnome-ish DOM projects, gdom and gdom2, but we were
discouraged from introducing any dependency on them into ATK, and 
at any rate they don't seem to have ABI/API backing outside of GNOME.

My understanding is that what we really need is something equivalent to
the w3c's reference implementation of DOM bindings for Java
(org.w3c.dom),
only in C.  If we have that or equivalent, we can return a "DOM" object
that
the rest of the world has a chance of either being able to import or
export.

We may in fact be wrong about the necessity for this but it appears to
be called for by the w3c "Web Accessibility Initiative's" user agent
guidelines.  
I consider it a matter for the future.

Regards,

Bill

>         Regards,
> 
>                 Michael.
>