So you’ve got a site and you want to say how great “standards” are, and “standards” now means that a certain way of developing things is better than another, equally standards-compliant way. Tables, they opine, are bad for layout, despite the fact that they are better and more consistently supported by browsers and accessible readers than relative or absolute CSS positioning of <div> tags.

How should you present this revelation to the widest audience possible?

Use tables for layout, of course.

Even better, you should brag about your redesign and how it validates.

Oops, it doesn’t validate to the loose standard you set for it.

Now they say that older pages won’t validate, but shemp on a shingle, shouldn’t the article where you brag about your standards-compliant redesign validate?

These are not errors introduced by commenters, but first-day-of-standards-compliance-class errors.

Now, check out the version of the Evolt article on the site of the author.

Yaaaay, he uses standards, and look, ma, no tables. Of course, his design is trivial, something that wouldn’t cause a talking browser for a blind person a startle if they’d done it in Netscape 4.x compliant-code. So in the case of a toy design, it works.

Do these people not code sites for clients based on designs given to them by designers who work in Photoshop? That’s the way things work in the real world, kids.

I have tried my damnedest over the past year to code table-free, or at least table-minimal, sites. At first my productivity plummeted. Sure, it’s a new thing, even though I’d been writing standards-compliant code since 1999. But this table-free thing I’d avoided. So I gave it a shot. So as I learned the tips and tricks–and there are more and more subtle tricks for CSS+divs than there are for table-based design–my productivity went down to about 20% above where it had been before when I used mainly tables for layout and CSS 1 for trivial, well-supported applications (mainly to replace font tags).

Still, every site requires me to grab another CSS+div coder at work and wrestle through issue after issue to achieve something it takes me far less time to do with tables. Some of them become Mozilla bugs that the coders of THE MOST STANDARDS-COMPLIANT BROWSER IN THE WORLD tell us are the fault of bad standards.

Now, my colleague usually shoots back, do I want to return to 1998, single-pixel gifs and font tags everywhere?

No, that’s a straw man, and a year ago, noted table-free hysteric Zeldman conceded this.

There’s no reason not to use CSS 1, non-box model stuff. Heck, for lots of things you can use the padding attribute safely. You can do some nice stuff to correct the problems with forms. You can definitely strip out all your font tags.

Every response I’ve read to Zeldman’s article above reads like “duh, that’s what we meant anyway,” or “well, you should just design to a standard, not to a browser.” However, Web sites are viewed on browsers, and if you think you can design to a standard and please your clients consistently, good luck. Even ESPN redesigned their site but put in a fair number of browser hacks and tested, tested, tested in browsers. In the end, you are designing for browsers. People may say “well, designing to a standard doesn’t mean not testing in browers,” but designing without a browser or two in mind means you’ll forever be re-implementing things to work around browser bugs. Because every last one of these sites defends some new practice by saying “it works in Mozilla.” If it didn’t work in Mozilla or any other browser, would they use it? They’re designing for Mozilla, not standards. And that leads to user-contempt and unemployment.

Quite simply, if the person paying you thousands of dollars to create a site for them uses Netscape 4 or IE 5 for Mac OS 9, telling them “your browser is teh suck, j00 are a l4m3R!!!” is a quick way to be fired.

I wish standards-compliant CSS positioning worked, I really do. float would be a wonderful property to use without testing it five thousand ways from Sunday because of the differing interpretations of standards and just plain bugs that exist. However, it isn’t, and though I may occasionally struggle to replace it, I have to concede defeat. I just can’t do it fast enough to make commercial sense. Maybe some of these guys who write articles on table-free design can do it, but since they point to their personal sites rather than working professional designs for paying clients, I sincerely doubt it.