Embittered generals

I had yesterday a quite long IM discussion that deserves a summary, and a reply, here.

We are "embittered generals", and we are a threat for the W3C and the people working for the W3C. We are trashing years of W3C evangelism for our own glory. We are also accused of playing, unconsciously, for Microsoft.

That's in substance one of the comments I had to face yesterday when I released my HTML Overlays proposal. Oh, and I also faced strong comments because, yes, I say my or our (because we're two on that, Laurent and I) proposal.

Who's that we by the way? That we gathers for instance the author of the current lines, the unmanageable fanatic Ian Hickson and the people of the WHATWG (but the critics said WTF), Tantek Çelik and the other XFN-maniacs Matthew Mullenweg and the well known I-don't-care-about-the-standards Eric Meyer.

We are evil. We are about to break the World Wide Web Consortium itself and send its employees to the Unemployment Office. Nothing less.

So what's the situation here? Let's give a closer look at what the XFN guys have done, what Laurent and I have done with HTML Overlays, and what the WHAT-WG does.

  1. The XFN people have used a technical possibility offered by the specification HTML 4.01 to extend the list of values of the rel/rev attibutes described in section 6.12 of the spec:

    " Authors may wish to define additional link types not described in this specification. If they do so, they should use a profile to cite the conventions used to define the link types. Please see the profile attribute of the HEAD element for more details "

    And that's exactly what they did. And because they believed others may like it too, they made a web site to talk about it. And others like it. So what's bad here? Ooh, those three geeks are evil, they made that spec (can it really be called a spec) alone?!? Can you believe it? And they only specified "positive" values for friendship, contact or love, nothing for ennemies, hate. They use english words, and some notions might be hard to internationalize. They use words that express geographical proximity, but nothing is said about the proximity itself. Urgh. Sure, sure, those three guys know a very little bit about standards, their code implementation and the Web in general, but is it really enough? And don't you think it's only self-glorification, that they used their already a bit famous names to get the eyes of the medias on them a bit more?

    Wait... HTML element names are in english, right? XHTML, SVG and other W3C dialects element names are in english, right? HTTP header field names are in english? And wait, did HTML4 offer to completely internationalize numbered list items? It offered a very restricted set of possible values, some of them completely underspecified. Did that make the World Wide Web collapse?

    Instead of congrats for a simple, backwards-compatible idea, or even a bad one but a bad one people seem to like, an idea people could  build a standard upon, and in the end a cooperation proposal, XFN authors faced harsh, and even insulting, critics from a few standards people.

  2. What Laurent Jouanneau and I did with HTML Overlays is absolutely similar. We used the possibility offered by the same section 6.12 of HTML 4.01 spec to add a new rel/rev value. So why are we so dangerous here? Here's a tentative list of reasons:
    • Just for the same reason Tantek, Matthew and Eric are dangerous with XFN: we did that in our garage, outside of the W3C. We did not talk about it with the HTML WG before release. Let me quote here Tim Berners-Lee in his book "Weaving the Web":

      " This mindset, I thought, was disappointingly different from the more (...) entrepreneurial attitude of developing something in the garage for fun and worrying about funding it when it worked "

    • We "extend" the browsers that can be extended. HTML Overlays are based on a JavaScript add-on. So, that's right, text-based browsers or people disabling JavaScript will see broken Web pages. Honestly, I think there are only two kinds of people who use non-JS aware browsers: on one hand, people who have no choice because of disabilities or a technical environment they cannot change, and these people deserve a solution to the accessibility problem we create using JavaScript; and on the other, people who do use JS-aware browsers and disable it because they think JS is the worst invention of all times. I have nothing to say to this second group, I sincerely don't care.
    • We did that because we do feel the HTML WG is on a wrong way with XHTML 2, because we think XHTML 2 is not innovative enough, is a bad compromise that takes ages to appear, and will take even longer to be implemented if it is ever implemented. We do believe, just like the XFN and the WHATWG folks, that the dozens of millions of HTML web pages are not going to disappear magically or switch to an hypothetic XHTML2, that the millions of HTML4 browsers are not going to go away tomorrow, that HTML4 is too old now, and that all non-intrusive paths  of improvement must be followed. The HTML WG has made a choice about HTML4: it is in strict errata mode, meaning that no addition is accepted. That's a stupid decision, but that's their decision. So we did our job with HTML4 in a way we were expressly allowed by the spec to use. And we cannot see anything wrong here.
    • We did that in one hour. Laurent expressed the the idea during lunch friday 27-aug, I spent ten minutes on it friday afternoon, an hour on it sunday night, wrote the paper monday morning. We really did that in a garage, and it works on 95% of the browsers in the world. If two crazy frogs can make their own extension to HTML4 that way, good or bad, what's the point spending a year in a Consortium and paying a $50k annual fee? What's the point spending months on XML include or merge mechanisms when two snail-eaters can design on a napkin about 150 lines (commented and not very compact) of cross-browser JavaScript code that provide a merge mechanism as powerful as XUL overlays?

    We are proud of the name of our company. We are Disruptive Innovations. Some take Disruptive negatively, and some think Innovations. Make your choice. Ours is made.

  3. The WHATWG was created by the W3C itself. I am not kidding. The constant refusal of the HTML WG to work on HTML 4 but for errata, and the recent Compound Documents Workshop could only lead to that result. The WHAT-WG people are not embittered, as I read yesterday. They do feel it's urgent to do something for the Web because the path followed by the W3C is wrong. And this "they" gathers implementors. They really regret the W3C is not willing to listen to their arguments, and regret that this work is not done around a W3C standardization table. But that's life. They are not going to drop all their beliefs about the future of the Web because someone said "no". They may succeed. Or fail. That's how it goes. The Web has always grown drinking the water coming from many sources. The only thing new here is that some people don't want it any longer.

I just don't want to comment on the self-glorification argument, that's too stupid.

About Microsoft, I am not fanatically against what Microsoft does. And I have no problem playing for Microsoft if I find what Microsoft does is good for the future of the Web.

We, and I am reusing the we detailed at the top of this article, are not working in stealth mode, not working on pure proprietary code. XFN is public, HTML Overlays are MPL'ed, WHAT-WG is working on a public mailing-list and releases drafts. In the long run, I am sure it's  browsers' adoption (or not) of these ideas that will say if they have to become a standard or not. There's not a single problem in XFN, HTML Overlays or WebForms that cannot be solved by people of good will. We hope it's clear we are of good will. Are you?

Update: comments now allowed. Flames/insults will be censored.


1. On Wednesday 1 September 2004, 09:34 by Anubis

Go on with your disruptive innovations and don't care about what the world says :)

2. On Wednesday 1 September 2004, 10:24 by jens.b

Just two words: Full ACK.

3. On Wednesday 1 September 2004, 10:54 by Pete

I get the feeling that some people overestimate their posiiton. Daniel and Laurent gave an example how to implement the XMLHTTPRequest DOM in a nice way.
That's nothing revolutionary. There are numerous approaches for that.
So this idea won't have any effect on the web.

Further I totally agree with Daniel that
- one cannot argue with people saying that JS is the worst thing on earth
- it's ok to cope with MS ideas like XMLHTTPRequest when they are helpful
- we shouldn't care if a HTML checker returns an error for a new link attribute

4. On Wednesday 1 September 2004, 11:26 by jmfayard

Just one question : have you tried to contact the WhatWG ? This is exactly what they search : new long-expected features which make the life of web developpers easier AND which are backward-compatible (clear migration path). If it's accepted, future browsers will perhaps be able to understand HTML overlays natively, and will not suffer from the accessibility problem caused by JavaScript.

5. On Wednesday 1 September 2004, 11:49 by lprevosto

My 2 cents : guess what ? Everybody has been doing this for a long time, especially for automatic HTML menu generation (you would like to be able to expand an item with content loaded from a server when you actually need it - this is lazy paradigm - and not having to load the all stuff at the begining).
But no, that's not that true : everybody would *like* to be able to do it.
Unfortunately, HTTP request methods in javascript are not standardized (XMLHTTPRequest in Mozilla, ActiveX in IE, ...)
I dont't know if we need another HTML or CSS add-on.
But we just need a standard way to fetch XML data from an HTTP server in javascript...

6. On Wednesday 1 September 2004, 12:24 by Tristan

Daniel wrote:

> I think there are only two kinds of people who use non-JS aware browsers

Errr... There are 4 of them:

* People with disabilities;
* People that do not like JS ;
* People that were told to disable JS and/or ActiveX because that, at the moment, it the only way to get rid of the numerous security exploits;
* _Google_. And Google is typically the kind of "user" you don't want to block.

For a first approach, your idea is very nice. Though, I don't think it can be used in its current form. Hence the reason why I used the words "(nice) hack" to describe it.

7. On Wednesday 1 September 2004, 16:14 by Bob Marcel

// whatever is $disruption

if (is_accepted_by_community($disruption))
$disruption->creator = "genius";
else $disruption->creator = "foo";

8. On Wednesday 1 September 2004, 20:36 by Bernd

Are there any URL's beside a IM chat that show the other side of that discussion. I mean that Javascript insult is already known and well documented. But the rest?

9. On Friday 3 September 2004, 18:32 by Tantek

Just a minor clarification.

Daniel wrote:

> XFN is public.

True. But more importantly, XFN <gmpg.org/xfn> is Creative Commons <creativecommons.org> licensed for all sorts of uses.