W3C

www.w3.org/TR/html5

It’s been a long time coming. Either 10 months (if you count back to when the current W3C HTML Working Group was chartered) or 10 years (if you consider when the HTML 4.0 Recommendation was published. Or maybe just 4 years (when Ian Hickson first blogged about developing a spec for a “Web Applications Language”.

Anyway, today we (the W3C and the W3C HTML Working Group) published a First Public Working Draft of the HTML5 specification. I lack the time right now to say much more about it (it’s either very late or very early here in Tokyo now), but to set a bit of context for this milestone, for now I just want to list of few random words and phrases and names from the months and years that have led up to it:

Cougar, validation, weblog, XML, draconian error handling, Mozilla, CSS, PHP, doctype switching, Google, collective wisdom of the Web, NGLayout, Gecko, XmlHttpRequest, Microsoft Internet Explorer, RSS, RFC 2616 (HTTP 1.1), Lars Knoll, KHTML, WebKit, ECMAScript 3rd Edition, feeds, XHTML, David Baron, Steven Pemberton, SVG, David Hyatt, Mark Pilgrim, HLink, Daniel Glazman, Jeffrey Zeldman, Safari, WordPress, Atom, XForms, Apple, Opera, Web Forms 2.0, Flickr, Presto, whatwg.org, Gmail, backwards compatibility, error handling, Google Maps, Ajax, Web 2.0, canvas, video, Gears

I guess that’s enough words for now.

5 thoughts on “www.w3.org/TR/html5

  1. I’m sorry I’m neither very good at speaking the English language, nor I’m experienced in mailing lists. Anyway, I would like to ask one question about the HTML 5 draft: The XHTML 2 Working Group planned to put the content of the meta element into itself instead of into the content attribute, which I think really makes sense. Unfortunately, the HTML 5 draft doesn’t adopt this idea. Does this decision have backwards compatibility reasons or isn’t it decided finally yet?

    1. Hi David,

      Indeed, browsers for a long time now recovers invalid html. One of the recovery mechanisms is to put back the “meta” element in the head when it has been misplaced in the body. HTML 5 being designed on the current behaviour of browsers, it is not possible to put a meta element in the body, or more exactly if people do, the element will be moved to the head section (in the browser model of the document.)

  2. Thanks for your answer, Karl! Unfortunately, as I expected, my English was not good enough to communicate what I really wanted to ask for. I will try to be more precise: I think the notation <meta name="asdf" content="xyz" /> is a bit complicated. The XHTML2 WG planned to simplify the notation to <meta property="asdf">xyz</meta>. In other words, the Content Model should be ( PCDATA | Text )* instead of Empty. What I want to ask is why you don’t adopt this idea to HTML5. Is it because you want to maximize the compatibility to current user agents and to pages written in earlier HTML versions?

    And I have another, quite similar but, indeed, more important question: The XHTML2 WG also planned to set the img tag’s Content Model to ( PCDATA | Text )* in order to extend the possibilities of providing alternative content for images. In my opinion, this is a great idea that should also be adopted by HTML5. Or are there also backwards compatibility reasons prohibiting that?

    1. The backward compatibility is the reason. Most of Web pages are invalid (around 95% and more). It means they do not respect any precise grammar. Very often they do not advertise the grammar they use (using a doctype at the top.) So browsers have developed techniques to recover this content. Put <meta property=”asdf”>xyz</meta> in a live dom viewer and see what is happening in different browsers.

      <!DOCTYPE html>
      <head>
      <title>boo</boo>
      </head>
      <body>
      <meta property="asdf">xyz</meta>
      </body>
      </html>

      This will create a specific DOM depending on the browser. In mine right now,

          * DOCTYPE: html
          * HTML
                o HEAD
                      + TITLE
                            # #text: boo
                      + #text:
                      + META property="asdf"
                o BODY
                      + #text: xyz 

      It means it puts the meta back in the head and leaves the text in the body, which was not the intent of the author. So not possible to adopt in HTML 5.

  3. Ah, I think now I see what you mean. Quite sad (especially the thing with the alternative (X)HTML for images as child of theirselves), but I fully understand your view. So again, Karl, many thanks for your help! I really appreciate it!
    (Oh, and by the way: I really like your XML 10 design, especially the header and the footer. Really good work!)

Comments are closed.