This is not exactly a very convincing article. For one the two examples do not display the same in FF3 (there's different amounts of white space between the three columns in the different versions), and of course, the examples looks completely different in IE 6 which is about 90% of the problem to begin with.
The articles I'm disagreeing with state that it's not possible, neither restrict that statement as purely an IE6 problem. The demonstration is to prove that it is possible, which it clearly is. I could tweak the positioning of the columns by one or two pixels but it wouldn't make any difference to my point, would it?
We Tablists* only have to point out one hole at a time to continue being right. Not restricting the statement just makes it harder to refute, since you have to cover all the ways in which CSS-for-layout is deficient, rather than just a single way.
* Table-for-layout-ists is just too long, don't you agree? ;)
You're actually proving the point those articles are agreeing on. Your CSS changed to make that possible. The point is when you change the ordering of DIVs your CSS often also has to change. So they are not as loosely coupled as one would hope a Model-View relationship might be.
