> I really wish xhtml would have won in place of HTML/CSS.

This makes me doubt that you understand what you are talking about. XHTML and HTML have exactly the same semantics. CSS has little to do with them (you can use CSS to style you XML documents if you wish so). With CSS3 and modern browsers you can pretty much avoid anything to HTML just for styling purposes. Heck, take a look at http://camendesign.com/ It's not visually complicated, but there are no IDs and CLASSes in the source.

The idea was that xhtml was supposed to work hand in hand with XSLT. Instead of HTML/CSS you would have XHTML/XSLT. You would process XML files with a turing-complete language to display nice styles. Every web site would be directly machine-readable. RSS would be completemy useless as they would be easy to extract from the raw XML.

I argued with one of my friends about exactly this point. He said that web development is great, that the HTML is semantics, CSS is style, javascript is everything active, everything neatly in their own files.

What I said is - look at ANY of top100 Alexa websites and show me ONE where style, semantics and action is not mangled together in one big chaos, where HTML is just semantics and not DIVs put inside each other in a hierarchy that doesn't follow semantic logic, but visual logic (with divs classes "UP", "upper_menu" etc), where javascript sits neatly in its file instead being all over the place.

I don't remember the outcome of the debate, because we then proceeded to talk about TV shows.

I suspect that is at least partly because of heavy performance optimization. Unfortunately, even the cleverest of abstractions tend to break or at least bend when you have to optimize something very heavily.

Because we all now that no display is as optimized as an html div box... They are currently using crazy optimization techniques in browsers to make javascript-heavy pages have the level of badly-designed native apps.

Same here. I think the return to "do whatever you want and we try to interpret it correctly" was wrong. Code in XHTML is way more readable, maintainable and easier to learn (IMO).

I still write my HTML5 in correct XHTML Syntax.

I freaked out a little when I started seeing do-whatever-you-want creep back in when I first read diveintohtml5.org (mirrored: http://diveintohtml5.info/).

Sure, you save some (or a great deal many) characters, but at the cost of making everything that needs to work with it more complex. (Browsers are a large slice of this pie, but by no means is the only slice.)

I'm going to sit back and wait for the cycle to come back around to XHTML-like correctness again.

"I'm going to sit back and wait for the cycle to come back around to XHTML-like correctness again."

Those who cannot remember the past are condemned to repeat it.

I was doing that, but then found out that the XHTML empty-tag <foo /> is not valid for all elements in HTML5.

See http://www.whatwg.org/specs/web-apps/current-work/multipage/... (via http://mathiasbynens.be/notes/minimal-html via http://news.ycombinator.com/item?id=3356389 )

I believe that is the first time I have ever seen anyone say XHTML would make something easier. (Just one example: http://mathiasbynens.be/notes/minimal-html )