Why are we not still using tables-for-layout?

I was chatting with a couple of long-time web folk today. As often happens when those of us who lived through the browser wars get together, the talk turned to old browsers, and old hacks to get round the issues in those browsers.

I was thinking through the timeframe of CSS becoming usable at all (with the launch of Internet Explorer 3 in 1996) and a small group of standards folks managing to convince a lot of the web that tables-for-layout were a very bad thing, and not to be used. The time between those two points was about five years. I launched my web development company in September of 2001, and the only time I used tables-for-layout in a client project was for HTML email. I had personal sites using CSS for layout long before that. Of course people were using nested tables and chopped up graphics for longer than that. However, there was acceptance in web design circles that a separation of content and presentation enabled by CSS was the future.

This happened despite the fact that by asking developers to stop using tables and instead use CSS, we asked them to simplify their designs dramatically. The early CSS for layout sites were two or three columns, and you only got a footer if you had found a way to create the Holy Grail.

Looking back this is astounding. It speaks to how small the group of “influencers” were at that time. The web, CSS, and open standards, owes a huge debt to that early group. To the layout examples at Glish, Blue Robot, and Owen Briggs with Little Boxes. To Jeffrey Zeldman and his A List Apart article showing us how to move away from table hacks. And, to people such as Molly Holzschlag, who died this week. Molly paved the way for web standards evangelists to follow, speaking eloquently about how we should perform our craft but also why we should follow these emerging best practices.