W3C home > Mailing lists > Public > www-html@w3.org > February 2003

Visual Markup (should HTML die?)

From: Todd O'Bryan <toddobryan@mac.com>
Date: Sat, 22 Feb 2003 17:53:54 -0500
To: www-html@w3.org
Message-Id: <84A57666-46B8-11D7-AFC0-0030656EC0E4@mac.com>

Maybe this is not the appropriate place to ask this question, but it's 
something I've been wondering about for a while, so I'll put it out 
there...

When HTML was first created, it was meant to provide a separation 
between content and presentation. We were all encouraged to use tags 
that indicated the function of our content rather than try to muss and 
fuss with getting the content to look a certain way. Of course, people 
almost immediately ignored this advice and created huge messes of 
loaded images, fixed the "optimal size" of their browser windows, 
ignored standards in favor of what looked good in their favorite 
browser, and generally mixed content and presentation to the point that 
Cascading Style Sheets were created to finally give people an incentive 
to really separate content and presentation.

Except that the separation between the two made the job of an HTML 
browser much more difficult, because now it needed to read two files 
and correlate the data between them, and, to my mind at least, also 
made the job of a site designer much more difficult because it's 
necessary to have two documents in mind whenever you create a site: the 
HTML and the CSS.

But, with the advent of XML, it seems to me that no page should ever be 
written in HTML as a first step. Because XML really allows you to focus 
on content rather than presentation, and XSLT makes it fairly painless 
to translate an XML document into whatever format your little heart 
desires, why are we holding onto the idea that something like HTML (an 
imperfect content-model with unnecessarily complex presentation 
support) should even exist?

Why don't we just dump HTML and create a VML, a simple, 
straightforward, easy to parse, quick to render, Visual Markup 
Language. The idea would be to create content in XML and then translate 
it to a platform-independent visual representation that would be more 
controllable than HTML but also faster to process since the browser 
would have to make fewer decisions. Some would say that, if that's what 
I want, I should use PDF. But PDF is cumbersome, slow to process, and 
doesn't have the interactive feel of a web page.

It seems to me this would be the best of both worlds. VML would say 
exactly what a well-formed document should look like, so browser 
incompatibilities and the nastiness of trying to get things to "look 
right" on multiple platforms would be solved, but people who think 
content should be preserved would be encouraged to use XML and create 
documents that would be far more content-rich in terms of markup than 
any HTML out there now.

Would people abuse the system and jump directly to VML? Absolutely. But 
probably no more than the people who butcher HTML now, and at least 
we'd be guaranteed that any VML-compatible browser would be able to 
view their presentation, no matter how lazy their attempts at content 
mark-up. (Of course, the simple ability to *completely* change the look 
and feel of a site just by re-writing the XSLT document might convince 
even the most impatient soul to do the right thing. If VML contained a 
basic graphics capability, this complete re-imaging of a site is not 
out of the realm of possibility, since much of what must be done with 
GIFs and JPEGs in current browsers could be graphical commands with a 
few inserted images.)

I know VML doesn't address the problem of accessibility, but the way 
HTML is commonly misused doesn't further that laudable goal either. At 
least, if people started with XML, they'd be more likely to provide the 
basic ability to access the content in other ways.

Maybe I've missed something entirely, and maybe there are good reasons 
to maintain HTML as the middle-man between content and visualization, 
but maybe we're just stuck with this thing because we started with it 
so many years ago.

Todd O'Bryan
Received on Saturday, 22 February 2003 17:59:20 GMT

This archive was generated by hypermail 2.2.0+W3C-0.50 : Tuesday, 27 March 2012 18:15:54 GMT