Re: [XHTML-role] How to define roles still needs clarification

On Sat, 2006-11-18 at 00:32 +0000, David Woolley wrote:

> Generally the specification of rendering has resulted in elements and
> attributes being abused to achieve that particular rendering.

I'd tend to blame that more on:

1. The inclusion of purely presentational elements in HTML.

2. The gap between the emergence of HTML and the emergence of CSS.

3. The (ongoing) misapplication of WYSIWYG to web editors.

> I doubt that most of the above elements will ever be important in the HTML
> that most people experience, which is primarily advertising in nature,
> nor in the main use within companies, which is as a mechanism for thin
> client form applications, and these are the real reasons that browsers
> don't support them.

As far as I know, most current browsers support INS, DEL, BLOCKQUOTE,
and Q to some degree, but only Gecko-based browsers and Amaya seem to
implement the CITE attribute. Users of other browsers can only access it
by applying add-ons.

My claim that failing to suggest possible implementations of the CITE
attribute contributed to its lack of support is based on discussions
with browser developers. The ones I'd talked to at first confused it
with the CITE element, which is widely implemented and which they
remembered for its suggested rendering. When I asked if they could
implement the CITE attribute, they agreed that they would like to do so
(indeed they often had existing bugs on the issue), but were unsure what
to do as the specification doesn't suggest any particular
implementation. But once I pointed to existing implementations as
examples, implementing it themselves began to seem a lot easy.

If you look at the beginning of the message I'm writing to you now, I'm
actually using the plain text equivalent of a BLOCKQUOTE with CITE. As
most internet users send emails, I dispute that such concepts are
unimportant to "most people".

Many internet users read, comment on, and write blogs. Blogging thrives
on explicit intertextuality. Easy generation and access to citation
information would make a lot of common blogging tasks much simpler. For
example, with a right-click option you could copy a selection from an
online source with its URI and paste it as a Q with CITE into your post
or comment. You could annotate the quotation with your own links, or
preserve links original to the selection. With a right-click option on
the Q, users could jump straight to the source. This isn't incredibly
difficult functionality to implement; I've been working on a couple of
Firefox extensions that will do precisely this.

Without such basic convenience tools, you see commentators adopting
crude citation mechanisms that are absurd in hypertext. For example, in
the longer comment threads, you'll often see commentators quoting
previous comments and referring to them by number (!), forcing readers
to scroll up the page to find the comment they're discussing.

This leaves aside the obvious utility of quotation and citation in
traditional journalism, academia, and law.

It may be that, like television, the web has become heavily
commercialized. But it does not follow that most people don't also use
the web as a source of information. The plural of anecdotes is not data,
but I don't know anybody who only uses the web for buying stuff, and I
do know people who would use the web for information and entertainment,
but would never buy over the web. For most people, most of the time, I
suspect advertizing, whether on television or on their monitor, is
something they learn to block and ignore, not what they are there to
see.

I do not believe browser developers will prove unwilling to implement
functionality for information-orientated hypertext, so long as W3C makes
the task easy for them. If browser developers are really the relentless
enemies of information-orientated hypertext you imply, then W3C, as the
custodian of the medium, needs to refocus its dissipated energies away
from creating new specifications, and towards improving Amaya.

In any case, my main query here is about how user agents are supposed to
provide access to new roles that their original developers didn't know
about. Whether such a role is "price", "slider", or "starmap", user
agents still need to know how it should render and behave.

--
Benjamin Hawkes-Lewis

Received on Saturday, 18 November 2006 15:27:45 UTC