Why can’t people do tableless design?
Jan 8th, 2007 by Karen
So I started to customize a piece of open source software that the Libraries are interested in using for our archival finding aids. I openned up the code and was immediately disheartened to discovered that the pages are laid out in tables. This isn’t such a problem on the main page. However, when you get into the findings aids themselves this results in nested tables. Anyone who has spent time looking at accessibility issues knows nested tables is BAD. Yet, again and again I encounter them on library webpages and systems. Five years ago this was less troubling because accessibility issues and using CSS for layout was just coming to the forefront. However, the fact that I’m still seeing this stuff now makes me furious. Are people living in retroville? Come on XHTML 2.0 is a working draft of the W3C. Yet we can’t get people to write compliant XHTML 1.0 strict. Sigh. Now what I thought was a small customization job has grown in proportion exponentially. Maybe when I’m done I’ll send the modified code back to the developers with a note.
Library application vendors are the worst. All of our major application vendors either use tables or so much javascript it could never be accessible.
A few years ago I tried pushing web standards but have just about given up. The people who decide on the look prefer “pretty” over functionality or accessibility. Then there’s our web editor who decided to put lines under headings by adding instead of adjusting the style sheet.
I’m not going to say it’s impossible to lay out the dsc without tables, but I think the component layout gets far too complicated (and the existing corpus of XSLT stylesheets use tables).
I will gladly incorporate any non-tabular layout you come up with into our archival finding aids system. I imagine most implementations would.
Hi Karen,
I share your frustration with antiquated table based web design. I actually spent a lot of time trying to develop an attractive table-less display for EADs. While it is possible, I ended up abandoning it in favor of a mostly table free design. I found that to get a fluid layout and cross-browser support tables were the best solution for the components list, also, the the components list seem to me to be acceptable as table data. You can see them in action at Princeton (http://diglib.princeton.edu/ead/), or on my dev site.
I started to use only divs as well, but what you have against tables I can’t understand. I didn’t find any advantages using divs, you can make a website using tables and still be W3C validated.
Doing both I can tell I like more the tables. My friend uses only divs, I digged into that to see where it all started. I think because I use fireworks(exports html with tables) and do most of pages writing the code while my friend uses Photoshop(exports html with divs). So graphic design programs decided for us :)
cheers
My problem with tables for layout is it isn’t semantic markup. The whole point of XHTML is to tag things as what they are not how you want them to look. Another good example of this is if you want to make a journal title or book title italics don’t use <i> or <em> this isn’t semantically correct. You should use something like <cite> instead and then style that tag (which by default is italicized) to appear as you want. Using non-semantic markup makes your web page much more difficult for machines (other than a web browser to read). For example, tables can make a page much more difficult for a screen reader to deal with particularly if your design has nested tables. If we want to get to a place where information can be pulled from a web page by different tools we have to use more semantic markup. Microformats are trying to get at this too. In the meantime, though that means I tag my pages semantically whenever I can.
Yes, tables should stick to their purpose: show tabular data only