Monday 07 October 2002


Today is Labor Day in Australia, a public holiday, so we’ve had a long weekend, which I’ve spent almost entirely in CSS Hell. A client asked me if it would be possible to create a CSS-only design for a commercial site, based on a layout created in Photoshop. “Sure,” I told him, “let’s give it a shot.”

We knocked up something quickly on Friday afternoon and the plan was for me to spend another couple of hours on Saturday sanding off the rough edges. The smoothing process stretched into Sunday, until I finally gave up. The design that looked great in IE6, Opera 6, and Mozilla/Netscape just didn’t quite work in IE5.

I’m not sure why I kept at it so long. Call me stubborn. I must like the feeling of banging my head against a brick wall, and the taste of the blood that’s trickled down from the gash in my forehead. Call me stupid, given that on Saturday afternoon I followed a link (from where, I can’t recall) and read this August 29 post of Zeldman’s:

More and more, we find ourselves creating transitional layouts that incorporate simplified table structures; use sophisticated CSS to add the kind of details that formerly required nested tables, spacer gifs, and other presentational hacks; and serve a basic style sheet to 4.0 browsers that approximates the display in modern ones.

We find that with these techniques we can create attractive sites that conserve bandwidth and look almost as good in Netscape 4.x as they do in modern browsers. We can’t achieve the same results using pure CSS methods.

Table layouts are harder to maintain and somewhat less forward compatible than CSS layouts. But the combination of simple tables, sophisticated CSS for modern browsers, and basic CSS for old ones has enabled us to produce marketable work that validates — work that is accessible in every sense of the word.

We spent today building a combined table and CSS design that works perfectly well. So I now understand why there are relatively few trouble-free CSS layouts and why they suit the simple requirements of most weblogs. I haven’t really lost my (CSS) faith and I appreciate Zeldman’s pragmatism—marketable work that validates and is accessible is a realistic benchmark for any commercial web design company. Still…

Call me disillusioned. This is probably a little unfair to Zeldman but it’s as though the Pope casually announced in St Peter’s Square one Wednesday morning: “Oh, by the way, did I mention I’m no longer a Catholic?”



Anyway that we can see the site you worked on?

As for Zeldman, I don't see his statement as one denouncing his 'faith' as much as it is a proclamation of tolerance. I think better of him for it.

Posted by: Burningbird on 8 October 2002 at 12:06 AM

Grrrrr. I hated Zeldman's quote the first time I read it, and now that I see it's going to be one of those oft-repeated quotes used as justification for all kinds of things, I hate it even more.

"Work that is accessible in every sense of the word" is such an incredible weasel phrase. It's like a philosophy freshman who is losing a philosophical argument and falls back to the "dictionary definition" of some technical term in order to make their point.

I'm becoming Stallman. I can just see it.

Posted by: Mark Pilgrim on 8 October 2002 at 12:16 AM

There's a place for pragmatism and their is a place for ideology. When dealing with a customer's money pragmatism is almost always going to be the wiser choice of the two paths.

A couple of things to keep in mind:

We are most likely always going to be in some sort of transitional phase between an old standard and the new standard, the old methods and the new methods. Professionals find acceptable ways to work through transitional periods.

Tables do have a place in HTML4 and XHTML documents. It is very likely that the layout you are looking at can be very loosely defined as tabular data.

Posted by: Justin Thyme on 8 October 2002 at 12:56 AM

You're not in danger of becoming Stallman, Mark. But you are in danger of becoming intolerant in your zeal. (Come to think of it, that does describe Stallman, doesn't it?)

Perhaps Zeldman was discussing accessibility from your perspective, but I assumed he meant it more from the perspective of validation.

Accessibility should be number one (and must check image tags at site again for alt attribute), but one can mix CSS and tables and still have a fully accessible site -- from your use of the word -- can't we?

Posted by: Burningbird on 8 October 2002 at 01:01 AM

Wait a minute, we're missing we're missing the important stuff here.

Did Jonathon just say "pulling a Loren"? Is this Bb's revenge for a past transgression?

Or, Jonathon, do you see this whole situation as Dorothea's revenge for male-pattern stubborness?

Personally, I'm so naive that I didn't even know it was unacceptable to use tables. They "look" so clean and eloquent.

Posted by: Loren on 8 October 2002 at 02:01 AM

Allow me to take a look at it Jonathan. If it is okay, send me the url by e-mail.
kristiaan at xs4all dot nl

On Zeldman's quote, I too felt a little bit betrayed when I first read it; later I felt better and realized that there was probably more to the story than I could grasp at that time.
No one should go out of their way just for standards and accessibility, ending up loosing the client or even loosing a business due to trying too hard but just not having been able to make that last hurdle.

I remember that Zeldman also said something in the sense of standards being there to be practical and not as a subject of religion.

Posted by: Kris on 8 October 2002 at 03:18 AM

Personally I consider Zeldman a politician, and a very good one. Which means that I always read what he says as being, in part, about positioning, in particular the positioning of Jeffrey Zeldman. What complicates matters is that he's sharp and knows a lot of sharp people.

I've had my own painful experiences with CSS, or rather with browser support of CSS, so I feel like I feel Jonathan's pain. On the positive side, support for CSS is improving and should continue to improve, which bodes well for the future. Meanwhile (and I hate to say this, Mark), I find myself in the middle somewhere, using CSS in much the way Zeldman described.

(Isn't it weird how these discussions start to sound like political discussions? I think this makes Mark a Marxist, or in this case a Markist.)

Posted by: Michael on 8 October 2002 at 10:23 AM

I agree with Z's pragmatism (using a mixture of tables + CSS), especially for the sort of mass market sites he's working on. Using tables for layout does not present major accessibility problems. See for example and .

What I object to is Z's use of the word "accessibility" to mean "backward compatibility". If backward compatibility with Netscape 4 is important to you, that's fine; there are resources available showing what CSS is Netscape-4-safe, and what accessibility features are Netscape-4-safe.

Web sites are not "accessible" to Netscape 4. Netscape users can upgrade their browser; blind users can not upgrade their eyes.

Posted by: Mark Pilgrim on 8 October 2002 at 01:06 PM

This discussion is now closed. My thanks to everyone who contributed.

© Copyright 2002-2003 Jonathon Delacour