First chapter of book HTML5: Up and Running is called "How Did We Get Here" and shortly tells story how HTML makes it to its current position. It was very exciting reading especially along with 17-years old mail conversation, which shows, how 'img' tag was created. The true is, there was 4 proposals, how images should be inlined to HTML, and some of them was perhaps even better than today used 'img', so what makes it to current form? It was just implementation of code in one of the that day's browsers. So in this case was true, that shipping code wins. Off course its not the golden rule, but it can help. Author in chapter 1 also introduces examples, when shipping the code just did not help. This was true for some of dead branches of HTML evolution tree.
Besides this examples author also gets over past evolution of HTML in W3C and WHAT working groups. As you can know, there was initiative from W3C to bury HTML and replace it with "better" and more strict XHTML. but unfortunately this initiative was not successful and even without support of W3C development of HTML continues. And then in 2007 W3C cancel its own initiative on really strict XHTML 2.0 and instead continues in development of HTML.
So where was the problem of XHTML? it was at most in lack of backwards compatibility and restrictions for developers. In version 1.0 XHTML has something called Apendix C, which said sth. like, obey these rules, but there is a way, how to make an exception. This Apendix C was removed in version 1.1 and did not make it to version 2.0, because work on this version was stopped. Without Appendix C all current sites would be needed to be rewritten to XHTML 1.1 and developers and users did not like it. Also there was an strict rule, which did not tolerate any errors in markups or parameters. So for example if you did not put " around parameter, the page just did not display, instead of tolerant browsers behaviour, which just can handle it somehow.
This "all or nothing" rule makes to to think little bit about broken and non-valid code on pages. Yes its true, that if we wont be strict on nice code, there will be more and more pages, but how to make some compromise? Because from developers point of view one has to optimize its site for different browsers in many cases, only because of browser dependant non-valid handling. So what about standardization of wrong code handling? Yes I know its very close to rendering engine and it will require big support from browsers site, but its kind of solution. Any other suggestions about this question is more than welcome.
