I’ve spoken about ‘considered harmful’ several times, and my favorite article (or anti-favorite, really) is still: Ian Hickson’s article about sending XHTML with a Content-Type: header of ‘text/html’.
That was bad enough, and horrifically damaging to advancing web standards – such as XHTML.
Now, however, the well-regarded and obviously well-educated and well-meaning gentleman has decided that XHTML is so terribly damaged by poor spec-writing and inconsistencies in implementation that he would create a new spec, HTML5. HTML5 is really a successor to HTML4, the last version of HTML we had before XHTML came into play.
This is an even worse step backwards than we had before.
XHTML, and XML, could have ushered in a lot of new technologies. Once enough content on the web was created in valid XML, high-speed parsers might be able to be developed which halt on invalid documents, thus causing higher-performing web-browsing experiences. Instead of serving HTML documents with tags like <html>, <head>, <body>, and others, you might get to the point where you serve documents like <invoice>, <lineitem>, <quantity>, and so on; styling these documents with XSL/XSLT. That could really be great, and usher in a world where you have one canonical format for your data, plus a stylesheet for human consumption. The computer can parse it, as-is. Wouldn’t that be great?
Yes, but hixie has ruined it for us. Because of his ‘concern’ about serving slightly dented HTML content to the unfortunate users of IE, all forward progress must stop. Good job.
Now we’re back to where we started. Thank you so much, Mr. Hickson. I tell you what, since you’re so concerned about validity, why don’t we just scrap HTML in total and make the internet be nothing but text documents. Instead of links, we can have instructions in the middle of our text such as, “To see Brady’s blog, type http://uberbrady.blogspot.com into the address bar of your text browser.” Then all documents will be valid, forever. OK? Great.