RE: Multi-channel content

David, 

Thanks for your thoughtful comments. I'm sure you are right about the differing 
experience of members of DIWG and CSS. By the way CSS forms a vital
part of the picture for DIWG and I should have mentioned it in my previous note.

It is worth saying that DIWG's approach is to extend the existing recommendations where
they are felt not to address the totality of device independence. So, for example, CSS 3 is
particularly interesting to the group, offering as it seems to the ability to extend
CSS in a regular way to include the additional features needed. Far from representing 
a radically different approach to device independence, DIWG's approach is to extend existing
specifications and approaches to try to provide capabilities that have been found useful in 
real multichannel implementations. It is worth noting that there are commercially available 
solutions based on the priniciples expressed by DIWG that support very large multichannel implementations today. These are based on combinations of XHTML, CSS, other standards and extensions.

Actually there is another important point here. A number of technologies produce potential solutions
to this problem. For DIWG, the issue is to provide technologies that make authoring affordable and help
to avoid the kind of fragmentation that occurred when WML was originally introduced. 

You raise a very good point about what constitutes a user agent. DIWG has considered this carefully
along with definitions of many other terms. Our glossary (http://www.w3.org/TR/di-gloss/) contains
the definitions used by DIWG.

I think I have to disagree on your view of the web. Surely it is at least the set of resources that
can be addressed via URLs that include the http: scheme? Indeed the curent draft of the Web
architecture document(http://www.w3.org/TR/webarch/) implies that it has larger scope than this and
covers other schemes too. Clearly not all of that material is necessarily covered by considerations
of device independence, but I think the web is more than just the HTML pages that constitute most 
current web sites. For example, there is a large and increasing amount of XML and RDF data accessible via HTTP to say nothing of other, different presentation markup languages that have been in use for a number of years, some of which are particularly important in specific geographies.

The aim of DIWG is to produce specifications that are extensions of existing W3C definitions and
that allow authors to create resources that can be delivered across multiple channels. DIWG's view
is that where the target device cannot natively use these resources, processors adapt the content
appropriately. Where it can, the materials, can be delivered directly to the user agent without
adaptation. This approach can support new devices that conform to specifications such as XHTML 2 and
client-side XForms etc. as well as providing support for the existing and legacy devices that do not.

As with other processors associated with W3C recommendations, (XSLT for example), DIWG says nothing 
about the implementation. The key is that the markup and other representations used for authoring are 
able to capture the intent of the author in a way sufficiently abstract to allow adaptation to
particular device characteristics to be possible, whether carried out in a browser, in some
adaptation processor or in some combination of both. XHTML 2, together with CSS, XForms etc. provide
almost all that is needed. DIWG's work aims to provide the balance.

Very best wishes and thanks for a stimulating discussion

Rhys



-----Original Message-----
From: L. David Baron [mailto:dbaron@dbaron.org]
Sent: 18 February 2004 09:48
To: www-di@w3.org
Subject: Re: Multi-channel content



On Wednesday 2004-02-18 09:02 -0000, Rhys Lewis wrote [text re-wrapped]:
> The very existence of the W3C Device Independence Working Group shows
> that there is broad industry support for the need for work to extend
> the capabilities of current web technologies to support the huge
> variety of devices that can now be used on the web.

It's worth noting, however, that the DI group was formed after some
existing, purely client-side, approaches to device independence already
existed (media-specific stylesheets) in HTML4 and CSS2.  This makes it
likely that the group is composed generally of those who are less
satisfied with such approaches, and thus likely that the participants in
this list would tend to be self-selected as those more against the views
expressed in the first message on this thread [1].

The CSS working group is continuing along the path of HTML4 and CSS2
with Media Queries [2], and it wouldn't surprise me if you found the
opinions on www-style [3] to be quite different, since the participants
there are self-selected for the opposite bias.

> On your comment about XHTML, while it is certainly possible to create
> XHTML documents that can be delivered to a range of XHTML compatible
> devices, this does not solve the problem of device independence.
> First, XHTML is not universally supported. XHTML documents are
> unusable on WML devices, for example.

This seems to beg the question of what the web is.  I certainly like to
think of the web as a collection of essentially universally-accessible
information in a small set of core data formats, and I find it hard to
think of a user-agent that can't access most of the information on the
web to be a web user-agent.  (If such a user-agent were accessing the
web through a proxy that transcoded "the web" into WML, the web user
agent in the system would be the combination of the proxy and the
device, and thus the web user agent in the system would support
HTML/XHTML.)

> others, CSS mobile profile. Some devices require stylesheets to be
> external, others require style definitions to be in line. And that is

Are these devices really web user agents, or do they live behind
transcoding proxies that are (in combination with the device) the real
web user agents?  How could they deal with the content on the web?

-David

[1] http://lists.w3.org/Archives/Public/www-di/2004Feb/0000.html
[2] http://www.w3.org/TR/css3-mediaqueries/
[3] http://lists.w3.org/Archives/Public/www-style/

-- 
L. David Baron                                <URL: http://dbaron.org/ >

Received on Wednesday, 18 February 2004 06:41:42 UTC