Slow professional suicide

Al Sparber makes perfect sense in article from June last year:

There’s nothing inherently wrong with using tables to layout a web page. They are great for rapidly deployment sites for clients who need to support a wide range of older browsers, or in any site that needs to display organized tabular data. Tables are not evil. But in terms of web page design, nothing is absolute. The target is constantly moving... but not in circles. To only use tables is as wrong a decision as to never use tables. Not to embrace CSS, could be tantamount to a slow professional suicide.

Web design is like medicine. Successful doctors are the ones who keep up with the latest techniques. You wouldn’t want to go to a doctor who doesn’t know the latest imaging and non-invasive surgical methods. I wouldn’t hire a web designer who couldn’t wax poetic about CSS, W3C Standards, Accessibility, and Usability.

Al has just started a very promising tutorial series on CSS Layouts: Woven with CSS: Quick Draw MacFly.

Posted 23rd February 2003 at 1:06 pm

3 comments

There remain two problems with pure-CSS multicolumn layouts: they *will* break at certain window and text size combinations, and right-navigation forces content down the page source. Try the QuickDraw McFly demo site. Shrink the window and increase the text size. The maincontent and sidebar divs first become incoherent and are next irresolvable. Now view the source. Alternative user agents (and search engines) must slog through the sidebar div before encountering the content. That's bad, in my view. Until CSS3 columns appear, tables (preferrably a single table) are the only reliable method for rendering multiple columns that overcomes these issues. If the first element in the <body> is a table and the main page content is contained in its top-left cell, accessibility is preserved, search engines are accommodated, and nearly unbreakable right-nav becomes a reality. Hey, I'm more than ready to jump to div-only layouts, and of course use them for single column pages. But they are not (yet) suitable for most multiple column layouts, a fact that's separate from whether one views tables as a cheat or not. Certainly it's good that more folks come to understand CSS-P; I don't mean to disparage P7 on this. Yet multicolumn div layouts only succeed under unspoken viewport assumptions that often don't hold, and I think that any recommendation to use them must take this up.
Lou Quillio - 23rd February 2003 15:23 - #
He's wrong. There is something inherently wrong with using table elements for layout - the fact that pages aren't tabular data.

What he means to say, I suspect, is that common symptoms can be alleviated. That's all very well, but the corner cases are often the most troubling thing about misuse of tables. How much more useful would the web be if semantic markup wasn't abused in this way?

Jim - 24th February 2003 11:40 - #
perhaps if tables were called layout then all these arguments would go away. After all the only real argument people have about a well crafted table layout site is that the tables are not tabular data.

I think w3c are going off down a poor route by not allowing some form of layout table structure. CSS3 still doesn't permit you to do everything you might want to with tables.

somthing nice might be css attributes that implement an 'aligment group' whereby I could give two or more different elements the following
1) alignleft:'myLabel' 2) alignRight:'myLabel'
this would ensure that the left edge of 1) would be aligned with the right edge of 2). Extend this to centers, vertical aligns, etc and you cover a lot of the things that css won't allow you to do which good graphic design asks for.

Tim Parkin - 24th February 2003 13:47 - #

Comments are closed.

Previously hosted at http://simon.incutio.com/archive/2003/02/23/slowProfessionalSuicide

Simon Willison’s Weblog

Slow professional suicide

3 comments