6 Aug 2010 11 am eastern

Earliest Web Doc is HTML5

Links and Anchors,” the very first document published on the web, is almost valid HTML5.

Hat tip: Jeremy Keith.

P.S. Got yours yet?

Filed under: HTML, HTML5, W3C, Web Design, Web Design History, Web Standards

15 Responses to “Earliest Web Doc is HTML5”

  1. Matt Latzke said on

    Besides not having a doctype, html tag, body tag, etc, sure!

  2. Mike said on

    We’re back where we started.

  3. Mason Stewart said on

    @Matt hahaha!

    @zeldman It’s kind of like wasting hours on tons of not-working code, and then realizing that you can accomplish your task with a single, well-written line. Bam.

  4. Travis Vocino said on

    …or Doc found another 1.21 gigawatts.

  5. pixelBender67 said on

    Love the new DOCTYPE

  6. Agos said on

    The fact that an almost-empty document is almost (!) valid is not really groundshaking.
    The document it links to, for example, is even less valid.

  7. Jeremy Keith said on

    Matt Latzke, the doctype *is* required (but only for legacy reasons) but HTML and BODY are not.

    This is not new to HTML5.

  8. Brade said on

    Also, is it okay in HTML5 not to quote your attributes? And is it okay for a piece of text not to have a tag around it at all? I’m genuinely curious…

  9. Martijn said on


    Also, is it okay in HTML5 not to quote your attributes?

    I believe it is.

    The validator link posted by Jeffrey defaults to using the “HTML5+ARIA, SVG 1.1 plus MathML 2.0 (experimental)” scheme which seems not to accept the non-quoted attributes but setting the scheme to HTML5 specifically stops this error from displaying.

    This means the only problem with the earliest web doc is the fact that a DOCTYPE is required.

  10. Bertil Wennergren said on

    “Warning: Using windows-1252 instead of the declared encoding iso-8859-1.”

    ??? The document seems to use ASCII and ASCII only.

  11. Jeremy Keith said on

    Brade, it is okay in every flavour of HTML (5 or otherwise) to not quote your attributes (assuming there aren’t any spaces in the attribute value). Only XHTML requires quoted attributes.

  12. Jeffrey Zeldman said on

    Allow me to clarify, as the point of this post may not be obvious to everyone.

    Mr Jeremy Keith observed that the first HTML document ever written was practically valid HTML5.

    In so observing, Mr Keith was not calling our attention to the extraordinary depth or magical foreknowledge of the first HMTL document (for it possesses neither).

    He was not saying that Tim Berners-Lee’s first, simple HTML page eerily presaged VIDEO and CANVAS or used advanced structural semantics. Tim Berners-Lee’s invention may have changed the world, but whatever his other gifts, Mr Berners-Lee is not a wizard or a time traveler.

    No, the point of Mr Keith’s observation and my echo of it here was quite different.

    The point was to show just how backward-compatible HTML5 is by design.

    HTML5’s creators, observing that XHTML 2 was not backward compatible by design, and also observing that XHTML 2 did not exactly catch fire with the public, had no intention of repeating the chief sin of the framers of XHTML 2.

    By design, HTML5 is intended to be backward compatible.

    Just how backward compatible is it?

    To find out, Jeremy Keith looked at the first HTML document ever written, and found that it was near-valid HTML5.

    In other words, the focus here should not be on the merits of the first HTML document. The focus should be on the merits of HTML5 and its promise to support everything that has come before while bringing reason, structure, semantics, and power to everything that will come hereafter.

    That’s the point.

  13. Peter Winnberg said on

    With the strong focus on backwards compatibility in HTML5 this is indeed an interesting find, however HTML5 has made several changes to the semantics of some elements that could make it incompatible with earlier standards. Changes like what the small element represents, how to mark up dialogue etc. could mean that it is not possible to just change the DOCTYPE and have a HTML5 document without manually reviewing the markup.

    Then again, HTML5 isn’t finished so these could just be bugs that will be fixed when time allows.

  14. Billee D. said on

    Exactly. That was also the point of my reply on Twitter the other day. The similarities are obvious and are there by design. The backwards compatibility of HTML5 with previous versions of HTML is hard to miss.

  15. Aaron said on

    I think utilizing practices such as quoted attributes, and keeping tags the same case makes readiblity and management a lot better. One can write bad looking code in C++ or Java, so I don’t see why HTML shouldn’t have that freedom either.

Comments off.