The Trouble With Web Standards

Another of the trends underlying the supposed death of the web is the shift from web standards to proprietary languages for developing networked applications. This is exemplified by the iPhone app ecosystem, which is inhabited largely by apps that could be websites but have been implemented instead using Apple's Cocoa Touch API. I touched on most of the justifications for taking this approach in my previous post. Basically, it was the shortest path for Apple to get to where they wanted to be in terms of features and tools. Conspiracy theorists might claim that they also wanted to maximize lock-in by making it harder to port apps to other platforms, doubtless citing Apple's decision to ban third-party development environments for iOS. There's probably some truth to that as well, despite the fact that they ended up relaxing these restrictions.

In many ways this echoes the long-standing antipathy of open web folks for Adobe Flash. Why use a closed tool with crappy accessibility, no view source and limited extensibility when you can bask in the limitless possibilities of the open web? The answer is the same: better features and tools.

The tension between open standards and proprietary alternatives is nothing new. Watching CORBA go down in flames in the 1990's while Microsoft built a massive developer community around COM had a profound effect on my attitude towards standards efforts. And we are likely to see it crop up again and again in future for the simple reason that both approaches have a lot going for them. Standards are terrific because they provide so much more leverage for a given development effort. My assertion that coding for the web is a "write once, run everywhere" proposition was poo-pooed in the comments to my last post. Fair enough, there are always some devilish details. But the fact remains that it's a darn sight easier to get an app coded using HTML and JavaScript to run on Windows, OS X and Linux than with pretty much any other approach.

At the same time, the often lethargic pace of progress in the standards world is a strong argument for forging your own path. For all the pie-in-the-sky rhetoric about the wonders of openness, the notion that standard bodies should be the primary driver of innovation in the software world is a design-by-committee vision of the web that no one in their right mind should subscribe to. To be fair, I suspect hardly anyone actually supports this view even if (like many technology folks, present company included) they aren't entirely in their right mind. But a lot of people come dangerously close when they bash Apple or Adobe's closed architectures or trumpet the "death of the web".

One of the most intriguing quotes in Wired's web-is-dead debate is Tim O'Reilly's assertion that:

Open and closed are in a great dance, always have been. Openness is where innovation happens; closedness is where value is captured.

There's a lot of truth to this, but it misses an important counterpoint. First of all, value capture drives innovation in its own right because the profit motive provides such a great incentive to take bold risks and pursue crazy dreams. Secondly, and more importantly in my view, open environments offer great opportunities for innovation within the confines of their particular parameters, but they make it very difficult to break out and create something entirely new and different. It took Flash to popularize video on the web, and the iPhone to popularize apps on the mobile phone because the open standards that coulda/woulda/shoulda done so simply weren't up to the task.

In an ideal world companies would instantly open up to standardization any successful invention. Then we'd get all the benefits of both private innovation and hundred-flowers-bloom open ecosystems. Unfortunately a lot of inventions would never leave the drawing board if this were the case. Better to accept that the most valuable standards often happen further down the road, after a particular invention has proven its worth and earned tons of money for its progenitor. In other words, Apple's iOS isn't a harbinger of the web's obsolescence. It's just business as usual in the technology world.

Matt

Matt