You're using IE. Scroll down.
home :: tech :: newsitedesign

The new site design

Redoing the site came about for two main reasons. One was me not being happy with the n-year old content in the old site and with the zombie book of days. The second, unsurprisingly, was me wanting to tinker with current web technologies.

The engine decision was pretty easy, i already used blosxom in the past and i pretty much like it for all the blosxomy reasons. The design itself was on the other hand excrutiating as i’m very much not a designer.

So it started as a technical exercise in xhtml/css which proved in itself pretty depressing as xhtml1.0/css2.0 isn’t particularly more interesting than html4.01 in terms of splitting content and design which again i like to think is the reason why all web designs are fixed pixel-acurate typographical designs (as opposed to the web designers being a bunch of dolts who can’t transition from paper to screen). In the end getting a nice multicolumn flowed design turned out pretty difficult both cause float doesn’t really work that well and flowed layout depends on the block order in the content side of things, not the layout side of things. And of course, i was cheating and ignoring the multitude of css bugs on ie6 to avoid utter madness.

In parallel it seemed a good idea to go utterly web2.0 and thick-client and push all non-static content to the browser. That was obviously a ploy to play with jQuery and newer versions of Dojo. So i had content being pulled from feeds, put into accordions and tabs all very much dynamic and moving on the browser.
And then it dawned on me it was all very daft. Nothing was actually being accomplished, i was creating huge load times, breaking box models left and right (hello microsoft!), fighting doms and nothing was accessible anyway. And in the end, most people would be viewing the content through a feed aggregator and would never touch the site itself. So that would be a wasted month if i didn’t come out of it much more experiences in the horrors of the web.

So in the end most of the stuff got scratched, all javascript went away, content got constructed server side and floats went away.
On the down side tables came back. Here xhtml/css is much better than html4. Instead of tables going into html they came back on css so we get divs that render as tables and we don’t break semantics but we can’t influence render order in css so we end up saying the table layout in html.

Eventually, 10 years from now, browsers will catch up to xml/css3 and i’ll be happy happy doing server side components that spew out xml that’s perfectly laid out by a designer i’ll wont need to know.

/tech | edited on 2008/05/06 -- permalink, click to comment
blog comments powered by Disqus
Archive: