In recent weeks I’ve noticed a burst of chatter about certain W3C standards, the working groups that define them, and the W3C itself. I have followed (and occasionally participated in) web standards discussions for several years, and I’ve been trying put this recent flurry of activity in context. I believe it can best be explained in terms of the Overton window.
Brief disclaimer: I work for IBM, which is a member of the W3C and is currently participating in several working groups within the W3C. I am not personally involved in any of these working groups, so I have no inside information on their progress. My purpose here is not to take sides for or against any particular standard or organization. I am trying to construct a framework for understanding the ongoing debate about standards and standards bodies. As always, my personal writing is not an official statement of IBM.
As I understand it, the Overton window is a visualization tool used by “think tanks” that want to sway public opinion on certain issues. You start by outlining the continuum of possible opinions on an issue, including opinions which seem ridiculous or unthinkable. Then you figure out the narrower range of opinions that people currently consider reasonable. This range is the Overton window. The job of the think tank is to move the Overton window in a certain direction, so that ideas that were once unthinkable become acceptable to discuss, and ideas that were once radical become popular and perhaps even become policy. Along the way, certain ideas that were once popular may “fall out of favor” and become taboo.
Of course, there aren’t any dedicated think tanks devoted to influencing opinions about web standards. But there are a lot of players — from multinational corporations to independent developers — who are interested in the future of such standards. And there is an obvious analogy between political ideas becoming policy and standards getting implemented and becoming commonplace. (In fact, some standards go beyond that and actually do become policy, like Section 508 mandates accessibility standards for US federal agencies.)
So what is the continuum of opinion about the W3C and web standards?
When I started blogging in 2001, the Overton window (the range of acceptable opinions) was squarely on #1 for every web standard you could think of. HTML, CSS, RDF, SVG, XForms, WCAG, you name it. Vendor-neutral web standards were all the rage, and the W3C was, quite simply, the only game in town for defining such standards. (We’ll conveniently skip over the orthogonal debate about whether there should be vendor-neutral standards in the first place.)
Fast forward to 2006.
The W3C is now under fire on several fronts, and various people have publicly criticized various W3C standards and (in some cases) publicly abandoned the W3C itself. For example, Joe Clark laid down his opposition to WCAG 2, the W3C’s next-generation accessibility standards, by arguing that WCAG 2 is not appropriate for any purpose (#7). But his entire article just sets the stage for his really radical move: a proposal to form a new group to define its own accessibility guidelines. Forming such a group could move the debate beyond #1, and producing concrete guidelines could move it beyond #2. Then the new group needs to argue for its own legitimacy (#3) and argue that their alternate guidelines are better on their merits (#4).
The reaction to the SVG Tiny 1.2 Recommendation has been even more pronounced. Some critics, such as Robert O’Callahan, argue that SVG Tiny is inappropriate for use on the public web and should be confined to mobile devices (#6). Others (Ian Hickson) argue that SVG is so technically deficient that it should not be used at all (#7). Still others (Björn Höhrmann) argue that the W3C has consistently violated its own process in ignoring formal objections and creating specifications which are incompatible with other W3C standards (#8, and possibly #9).
Some of these critics advocate a competing standard, the
<canvas> element, which provides some of the capabilities as SVG (albeit approaching the problem from a completely different perspective). The
<canvas> element is part of the WHATWG’s Web Applications 1.0 specification and already has partial support in some browsers. So I would say that the Overton window for SVG-vs-canvas is currently centered around #3-5. Those in the anti-SVG and anti-W3C camps are trying to shift the Overton window past #3 (by building out implementations of
<canvas>, which could confer legitimacy on the WHATWG as a whole), past #4 (by explaining why
<canvas> is better or easier or faster or whatever), and into the #5-6 range. It will be interesting to see if this shift occurs in the next 6-12 months.
On the flip side, what can pro-W3C advocates do to shift the Overton window in the other direction? First, they need to honestly acknowledge where the window is today for particular standards like WCAG, SVG, or XForms. For example, if my analysis is correct and SVG is currently hovering in the #3-5 range, then one tactic would be to argue that supporting both SVG and
<canvas> would be prohibitively expensive or unwieldy (shifting the debate away from #5). Another possibility would be to demonstrate that
<canvas> is technically deficient or inappropriate (shifting away from #4 and towards #2). Another would be to work towards getting the WHATWG itself to disband (shifting away from #3 and towards #1). Again, it will be interesting to see if this shift occurs in the near future.
Finally, it should be noted that there is usually no benefit in arguing too far outside the Overton window. People who have been claiming since 2001 that the W3C should be disbanded (#9) have had a negligible impact on the ongoing debate. Their opinions are just too far outside the range of acceptable opinion, and they risk being dismissed as kooks. If the Overton window does eventually shift far enough that such opinions become acceptable, those people may feel vindicated, but it is unlikely that they will have been responsible for shifting the debate.
For what it’s worth. <canvas> and SVG are not competing “standards.” They both fulfill a different purpose.
— Anne van Kesteren
2 and 4 are odd positions in one way, because both revolve around the supposed technological superiority of W3C’s choices. But even where this is not a factor a standard is still desirable. It would be like the rules of a game: they are arbitrary and could just as well have been something else, but simply because they are agreed upon interaction is possible. Isn’t this why W3C was riding high a few years ago? Microsoft and Netscape implemented whatever “extensions” to HTML code either wanted, and people saw what the result of that was: that the user needed to have a particular company’s software to view a webpage as it was intended to be seen. It kills inter-operability.
In 2001, most of the Web *design* community’s overton window was squarely on the position that the only relevant standard is what the latest version of Internet Explorer supports; all else is gravy.
Things have drastically changed already.
— Jeff Read
Anne: a 2-minute Google search pulled up plenty of examples of people talking about, presenting on, and demonstrating SVG and canvas as if they were competing standards.
To say “well they serve different purposes and people should support both” is just to reiterate #5 in my continuum list. That doesn’t contradict my argument; it validates it.
dbaron’s post doesn’t seem to fit into your argument. Or, maybe it’s a combination of #1 and #9.
— Robert Sayre
My article about RDFa and microformats does fall somewhere around your #5, but I never intended a value judgement in it; merely a statement of fact. There simply _are_ other formal and informal Web “standards” groups, for suitable values of “standard”, and implementors have little choice whether or not to use them.
— Evan Prodromou
This is bad news as far as I can see. What impetus is there for John Smith/Joe Bloggs to bother making his site validate (as per the W3C validator) or accessible (as per webXACT now or any future standard it checks for, i.e WCAG 2) if none of the gurus agree that is the correct way to go.
As bad as a standard may be, it’s a standard that everyone can work towards. Once people are there, then work to change that standard. By abandoning something just because you disagree with it aren’t we just heading back towards the accessibility equivalent of “Optimised for Internet Explorer, 1024×768 resolution”
I think dbaron’s argument is a kind-of variant on #5 but without the requirement that the W3C itself be internally consistent. The W3C merely has to produce standards and, instead of having an overarching plan, “let the market decide” which standards should be adopted in which environments.
The response of the W3C to this seems to be to try and absorb standards developed on the outside e.g. Web Forms 2, presumably with a view to reasserting position #1.
This is a very clean and thoughtful assessment of the situation, Mark. Thank you. We have discussed in in the SVG WG, in the context of the larger controversy surrounding SVG. There probably are people in the W3C who want to take the tactic you describe. I must confess, though, that I don’t personally agree with that approach. I don’t think that the W3C should try to wipe out the “competition”. I think it should work to improve its relationship with implementors and authors, such that there is neither a real nor perceived disconnect between the goals of the W3C, the authors, and the implementors.
You really should not have let out the secret of the Overton window. Now it is likely to find its way into the mainstream consciousness, thus limiting its effectiveness. You will be receiving a visit from the Fellows shortly.
— Joel Dueck
Your article reads as if W3C was losing its power to define standards. I never felt that way, and I am perfectly aware of the “competing standards” you have mentioned. W3C defines the framework (XML, CSS etc.) and others build applications on top of it. And with applications I don’t mean binary code, but DTDs and XML Schemas. Consider DocBook and XML-RPC, they have both become de facto standards without going through the W3C organization.
In short, I believe this evolution is *by design* and absolutely intentional.
— Daniel Lorch
I really liked this article, the information about overton window and how it is used to change common opinion over time was excellent.
— Kenneth Power
W3C validation? Who cares? Still Microsoft IE validation rules the net :(
Mark, how do you have time to work and have a life and still stay knowledgeable about all these things. Even without having a life, I still can not keep up.
Si: Yes, having multiple standards organizations in the same arena dilutes the authority of all of them. The people setting up alternative standards organizations know this. That they are going ahead anyway may indicate recklessness, but more likely it indicates the severity of the W3C‘s problems.
I am no longer accepting public comments on this post, but you can use this form to contact me privately. (Your message will not be published.)