(Cross posted from dev.aol.com)
Aside from us hapless Web developers, few people really think about HTML when they surf the Web. The average user’s Web site of choice is likely to work with their browser of choice. The fact that this is the case can largely be seen as a testament to the W3C standardization process — HTML just about works, tag soup notwithstanding.
Not many people know about the often acrimonious debates between Microsoft and Netscape that induced this kind of interoperability in the past, or about a certain HTML working group meeting in Colorado a few years ago where everyone worked all night on HTML so that they could go skiing all day. The point is, HTML is out there — the mavens have spoken. So what’s all the fuss about?
This discussion is all about the new direction the HTML Working Group intends to go. For a long time, the W3C’s HTML activity focused on XHTML2, which introduced incompatibilities with previous versions of (X)HTML. No desktop browser that the masses use supports XHTML2, and conversely, no AOL (my employers) or Time Inc. Interactive web property uses XHTML2. W3C’s focus on XHTML2 alienated browser vendors and the Web development community, and thus a new group — the WHATWG — energetically took on the task of crafting the future direction for HTML.
At a Technical Plenary in Mandelieu, France, in 2006, AOL and a few browser vendors noted our common agenda for the evolution of HTML. Essentially, mere document semantics weren’t what Web developers were clamoring for from their markup, and we wanted something that was at least cognizant of the work the WHATWG was doing, especially in maintaining compatibility. The W3C has now decided to pay attention again, and has proposed a few new working groups. Tim Berners-Lee, Director of the W3C, blogged about this in a post that generated a lot of optimism at AOL. Here are the proposed working groups:
Apple’s Safari guys made their views on the HTML Charter public; so did Daniel Glazman, developer of Nvu, who raised some concerns, including about having Chris Wilson as chair; Chris Wilson responded (and then there was some back and forth between Wilson and Glazman).
Now, it wouldn’t be far fetched to say that big content companies such as AOL LLC (a subdivision of Time Warner) are all about (X)HTML. The Web is our publishing arena, and is what keeps us in business. What do such companies have to say about the future direction of the content model that is their bread and butter? As AOL’s AC Representative to the W3C, I was tasked with coming up with a response to the rechartering. While I won’t share every nuance of our response publicly, here are some larger talking points that I feel comfortable sharing.
I’ll start with one of the contentious issues in the blogosphere — that of the chair of the working group. Should a browser vendor such as Microsoft be allowed to chair the working group? AOL owns Netscape (and its browser legacy), as well Windows software that embeds IE. I was once Netscape Communications’ Technology Evangelist, propagating the Gecko layout engine as part of the Mozilla project. At the time, I had little “geek empathy” for IE, particularly since one of my team’s aspirations was to see Gecko embedded in the AOL flagship product, the Windows AOL Client. That didn’t happen. And as a personally agitating coda to this, a VP at AOL blogged articulately and intelligently about why IE is a substantially better choice for AOL’s desktop software on Windows.
I say this without excessive rancor (given my past), but AOL builds a lot of Web applications, and they’ve absolutely got to work on IE for widespread adoption. IE is, after all, the dominant browser in the marketplace. Our Web developers are often held back because of standards support issues in IE, and are clamoring for it to do things that other browsers do. If Microsoft is genuinely interested in the evolution of the Web’s content model, alongside their proprietary developments, then we’re all for it. IE7 is evidence of progress. We have no objection to a Microsoft chair. Standards are not made in a vacuum, and are NOT the monopoly of one company alone (who said that?). As chair, Microsoft is obliged to lend order to the riotous (and hopefully productive) discussions. Bring it on!
And for what it’s worth, Chris Wilson will probably be good at it. Sure, there’s room for a co-chair, and I said so in AOL’s official commentary to the W3C.
Another observation is that the HTML WG doesn’t plan to explicitly liaise with external groups. I’m gathering that for certain Member-specific patent policy purposes, the liaison with the WHATWG is not made explicit, even though the WHATWG’s (X)HTML5 does an enormous amount of the heavy lifting when it comes time to consider the future of HTML. There have also been some implementations that have come out of WHATWG work. It stands to reason that a lot of the future HTML WG’s work ought to be done cognizant of the WHATWG, whether or not an explicit declaration for liaison is made in the charter.
Another aspect of not having external liaisons in the HTML WG is that it seems that only the XHTML2 WG will have coordination with OMA, and appears to be working on XHTML Basic for convergence with XHTML Mobile Profile. That’s fine, but from a standardization perspective, having an easy-to-understand story for mobility is important, especially given a new HTML activity. Importantly, we worry that there will be a fractured XHTML landscape. The WHATWG folks refer to the XHTML5 specification; the W3C has released XHTML 1.0, XHTML 1.1 (Modularization), and XHTML2 (Draft). And, the new HTML WG will also have an XML serialization of the markup. That’s a lot of XHTML specifications!
We also raised the issue of extensibility. Currently, you can extend HTML using various procedural mechanisms. The Atom specification has a clean extensibility story via namespaces, with prose that describes the parse model. In general, I think that giving some thought to a declarative extensibility model for HTML is probably a good thing, particularly given the possibility of diverse user agents that don’t have to worry about legacy markup, and can’t rely on procedural mechanisms. I wonder if the future HTML specification will even leverage DTDs, or go with prose and a handy RELAX NG schema, like Atom does.
Other observations include accessibility; the WAI-ARIA activity favors the use of the XHTML2 @role attribute, but XHTML2 and HTML now live in separate working groups. We’d urge intelligent coordination and specification development with accessibility in mind. The HTML WG charter calls for liaisons with the Web Accessibility Initiative, which may be a sufficient condition to encourage everyone to work it out. Early prototypes using the XHTML2 @role attribute look very promising, and I’d wager that content companies like ours are always going to be interested in technology that improves accessibility.
AOL is not going to participate in the XHTML2 working group, since we don’t use XHTML2 anywhere. We recognize that it is useful to some people, and support the charter as is. We also support the XForms charter, and didn’t raise any commentary that is worth noting about it.
As for the ongoing discussions about the feasibility of deliverables and of testing and success criterion, we decide to leave that to browser vendors. As a content company (and as the aforementioned hapless Web developers), we feel that this is an exciting development — it’s about time the standardization process for HTML came to life again. It’s about time we all worked together to realize the full potential of the Web, without breaking backwards compatibility.