Re: CSS validator bug - box model hack

I'm heartened, but not surprised, by the mainly positive response 
from respected colleagues at the W3C to whom I've written about this 
issue privately, and delighted by Sijtsche Smeman's reported quick 
effort to fix the bug.

I'm surprised that Kynn, of all people, would accuse me and other web 
designers and developers who promote W3C standards to our colleagues, 
and who sell standards-based work to paying clients, of "elitism," 
"scorn" for people who use the web, or "hatred" for the W3C.

Kynn was there when some colleagues and I founded The Web Standards 
Project in 1998, and he participated in our earliest conversations 
about how to convince browser makers to fully support W3C specs.

As I wrote to a friend in the W3C's QA group, we've managed to 
persuade many people who never gave the issue any previous thought 
that it is essential to author to spec and to validate their work. 
This is why you now see so many badges and validation links on 
commercial sites, public sector sites, and personal sites all over 
the web.

It is also one reason why many client RFPs now include accessibility, 
structural XHTML markup, and CSS layout among their requirements.

When CSS layouts stop validating even though CSS2 says they are 
valid, it sends a message that undermines what we've led our 
colleagues and clients to believe.

"You see," some will say. "These so-called standards are always 
changing. They can't be trusted any more than browsers can be 
trusted. We might as well do everything in Flash/ build for IE 
Windows only/ use table layouts" etc.

Now, that's the wrong message to take away from a bug in the 
validation service, but there are people will take that incorrect 
message to heart, undoing the progress we've made.

That's why I believe it's essential to fix validator bugs promptly. 
The validation service is, in many ways, the only part of the W3C 
with which many site designers, developers, and owners interact on a 
daily basis.

Then, too, if many of us continually hector browser makers and tool 
developers or complain about poor conformance in IE/Win or any other 
browser, it would be hypocritical and bizarre not to ask the W3C to 
fix conformance problems in its tools -- particularly when some of 
these problems were reported over two years ago.

I appreciate that these are open-source tools built by volunteers, 
and do what I can to encourage coders in our community to pitch in 
and help. (There's a message to that effect on my personal site right 
now.)

I mentioned the issue of membership fees, not out of "hatred" for the 
W3C, but simply to ask the members of this respected organization to 
consider setting aside a small budget to keep the validation services 
in good working order and fix problems fast. We're all familiar with 
open source successes, which can be inspiring and can transform the 
market. But we've also seen a market-leading web browser disappear, 
at least in part, because it took four years for the underlying 
open-source code to come to market.

A single membership fee, if set aside as validation service budget, 
would cover the cost of putting a single, dedicated coder on retainer 
-- that was my point.

Having raised it, I leave it to my colleagues at the W3C to decide 
whether or not the point has  merit.

As to the Box Model Hack, I don't claim it is the only or best means 
of ensuring that CSS layouts can be made to work in older, less 
compliant versions of Internet Explorer used by millions of 
consumers. I simply state that it's out there and it's valid per 
CSS2, hence the validator should return a warning instead of an 
error. (Sijtsche and other W3C members seem to agree.)

The Box Model Hack was created by a member of the W3C who actively 
contributes to the CSS working group -- hardly someone who "hates" 
the W3C. The year it was created, it enabled designers to stop 
worrying about IE5/Win's compliance problems and begin doing CSS 
layout in earnest. In so doing, it promoted the W3C's efforts to move 
web builders away from old-school non-semantic table layouts and 
toward the future.

Is it time to retire the Box Model Hack? Should it never have been 
used at all? Intelligent, informed people may argue either side of 
that coin. But even if you come down on the side of not using the 
hack, a bug in the validation service is not the right reason to walk 
away from it.

I'll ignore any other comments made by my colleague in his note about 
whining, elitist, W3C haters, because such comments are just wrong, 
as anyone familiar with the issues and players will know.

jeffrey

-- 

http://www.zeldman.com/dwws/  	Designing With Web Standards
http://www.zeldman.com/      	Daily web design news
http://www.happycog.com/     	Web design, consulting, and publishing
http://www.alistapart.com/  	From pixels to prose, coding to content

Received on Monday, 9 February 2004 10:45:18 UTC