[development] Default core theme
Thomas Barregren
thomas at webbredaktoren.se
Wed Oct 25 11:01:45 UTC 2006
Morbus Iff skrev:
>> Well you can't have your cake and eat it too... people want tableless
>> themes. They are impossible without at least some degree of CSS
>> hacks. Otherwise, you must go for tabled, and hardcode it in the
>> markup, which
>
> But you'll have to admit that there are CSS hacks that are more
> "acceptable" than others. Accepting a 20000px padding hack is far more
> difficult than accepting a html > sorta hack (or whatever the hell
> they are - I tend not to use any CSS hacks for my themes, but then
> again, I don't build three column themes every day).
I must object! The 20000px padding solution is definitely awful, but
that doesn't make it defective or inappropriate. On the contrary, the
20000px padding is a perfect valid solution to accomplish equal length
columns in CSS. And more important, it works in IE6, as opposed to the
otherwise preferable way to do it: position: absolute, top: 0; bottom: 0.
The actual stumbling-block for One True Layout
<http://www.positioniseverything.net/articles/onetruelayout/> is the
problem with linking to anchors in elements within the containing block
<http://www.positioniseverything.net/articles/onetruelayout/appendix/equalheightproblems>
due to a bug in the CSS 2.1 specification itself. If we can accept that
the default core theme only works with modern browsers (Internet
Explorer 6 and 7, Firefox, Safari, Opera 9 and others), a solution would
be to use the preferable solution above, and feed IE6 with 20000px
padding solution without the overflow: hidden on the containing wrapper
element.
> Then again, I'm of the position that a properly built tabled layout
> CAN be semantic. That's a long ass discussion for another day though ;)
There is no need for "a long ass discussion". It follows immediately by
simple logic that you are wrong. :-) By definition <table> should only
be used for tables. So therefore, if you use <table> for other purposes,
it isn't semantic. By /modus tollens/ it follows that a table based
layout can't be semantic correct. If you claim otherwise, please tell me
where the logic fall short.
Regards,
Thomas
More information about the development
mailing list