webprutser - 11:35 am on May 20, 2011 (gmt 0)
Thanks for the new input, tedster and rocknbil.
I've done some reading in the "historical posts". A lot of mixing up of tables and css, as if those can't be combined, but I read it as "positioning by css" and it seems I'm not exactly the only one who struggles with that part of css.
When I first read the theory of it, it was called the boxmodel, it sounded all very logic to me, it still does. Why it turns out to be so hard in practice I don't know, it seems very odd that so many people who have experience with websites yet and know how to shape things with tables get stuck on positioning with div's. I tend to say "if you make a tool so hard to use, I'm not interested in using it as long as I have a good alternative (this discussion of course, is about how good this alternative really is). First improve your tool, than I'll give it a try."
I must admit I don't know where we are with CSS3 at the moment and wonder if anyone knows if that has improved things from this perspective. Something else I read - and read before - is XML, but I'm too inexperienced to see if it would make sense to put my time on that track.
The discussions gave me some arguments to consider table-less design: page readers, loading times and mobile phones (on that last one I read different points of views, but that's a matter of testing my own website).
It also gave me arguments to don't make an entire switch to table-less sites.
The best of both worlds would also be an option to me. Make one table with as less tr's and td's as I can, only for the main positioning. If I get rid of heights and widths as well, I have no more styling info in my html, it would all be in the CSS-file.