Apparently, HTML5 hasn't "launched" yet.

You might remember a few months ago when the Arcade Fire released an interactive multimedia experience called "The Wilderness Downtown." The conversation about the piece was two-fold: the general public enjoyed the content, while the tech-savvy community was abuzz with the fact that the project had been built in HTML5.

(But the spec isn't finished, guys. It's not "ready.")

I spent this morning's first session at a panel on how to build cross-platform applications using HTML5. Using technologies that we traditionally consider as web-only, it's now possible to build applications that can be quickly and cheaply ported to both mobile and desktop platforms as applications that leverage the native features of the OS.  As a developer, this is exciting stuff, and it has made me want to rush home and start immediately tinkering with some of the tools suggested by the presenter.

(But HTML5 isn't "done," so. . .)

Okay - continuing the joke is a fairly silly conceit, but I think the point is clear: the genie is out of the bottle. In my second panel this morning, a couple of the people involved in creating the HTML5 spec discussed some of the political wars being waged behind the scenes during the development process. Concerns about accessibility, digital rights management, and artifacts of previous specs are all part of the set of issues that are delaying the official "blessing" of the W3C's definition of HTML5.  On one hand, it's hard to hear the arguments and not consider them to be valid and important concerns, but at the same time, how much longer can the spec be delayed before these discussions are made moot by the fact that an unfinished spec has already become the de facto standard? 

If the issues are debated for much longer, the ship will have already set sail (if it hasn't already).  One of the panelists said a scary thing while describing the concept of web-sockets (a powerful HTML5 feature that allows for what is essentially bi-directional http). Currently, that portion of the spec is in version 5. The panelist stated that developing for the version 1 spec of a year ago would've been like creating "another ie6.". If that analogy doesn't strike fear in the hearts of developers, I'm not really sure what does. The web-socket spec isn't even expected to be stable until version 6. HTML5 is already being used, but segments of it are still half-baked and flawed, and there is a very real danger of creating another "box-model" problem, where different browsers and mobile platforms interpret the spec in different ways, or render content in ways based on previous, unfinished versions of the specification.

I think the community has learned from their past mistakes. The hundreds of people involved in developing the HTML5 spec come from all areas of the development community, and all of them have an interest in not replicating the cross-browser mistakes we have already suffered through. Still, all of this careful planning can't fully account for the fact that people are excited about this technology, and they are rushing to use it, regardless of its preliminary nature.  There are already enough examples of its use that I was able to attend a panel debating HTML5 as a flash-killer.

One of my favorite parts of SXSW has been the way it sparks these intensely geeky conversations, centered around the kind of minutiae that aren't often discussed in such detail. To some extent, talking about the politics of a web spec could be seen as debating an already dead-issue, or posing a purely rhetorical question. At the same time, in talking about the process, we can anticipate where the points of friction are likely to arise, and hopefully outmaneuver some of the pitfalls. I know that if I start experimenting with HTML5 in the near future, I'll be conscientious of web-sockets and other new features of the spec, and I will also be wary of suggesting any radical shifts in technology until I feel confident that we're building on a stable foundation. Yes, the new spec is extremely exciting, but it's probably worth using a little caution until it's a little less raw.

Labels: front end development, HTML5, sxsw