Toolkits for user innovation

Technology trends obey certain predictable laws, among them that every action produces an equal and opposite reaction. So the recent backlash directed at AJAX and other "Web 2.0" technologies was no surprise.

Nicholas Carr called Web 2.0 an "ecstatic vision" expressed in "the language of rapture." Joel Spolsky called it "a big, vague, nebulous cloud of pure architectural nothingness." The infamous Web 2.0 meme map was brutally lampooned on a blog with the tag line "Please, God, Just One More Bubble!" Tim O'Reilly, co-producer of the second annual Web 2.0 conference that triggered the backlash, admitted that even he was worried that the hype was getting out of control.

Hype serves nobody's long-term interests. If we scrape away the ecstasy and rapture, can we find sound principles? Yes, but some of them need to escape the orbit of the O'Reilly alpha-geek subculture and colonize the wider IT universe. In particular, the notion of "hackability" -- a prime article of faith in the subculture -- is in dire need of a makeover.

In a recent New Hampshire Public Radio story about my flood screencast, NHPR's Jon Greenberg said I had hacked into Google Maps to add GPS coordinates. I winced when I heard that because, though I'd used the word myself, it sounded awful in a mainstream context.

I've made the same mistake before. In a blog essay about AJAX, I used two Google searches to compare the level of development activity surrounding Google Maps and TerraServer. A search for "google maps hacks" returned 3,430 results. A search for "terraserver hacks" returned one. My former Byte colleague John Montgomery, now a Microsoft executive, questioned my use of the word "hacks" as a metric of success. "A hack, by its very nature, is a workaround," he blogged. That connotation doesn't necessarily apply in the alpha-geek subculture, but John was right. If there's a principled argument to be made here, it must fly a different flag.

How about Democratizing Innovation? That's the title of a recent book by MIT's Eric von Hippel , who notes that users of products and services -- and by users he means both individuals and companies -- often innovate on their own rather than relying on manufacturers to do it for them. And not just in the realm of IT; a survey of employees in 74 pipe-hanger installation companies found that 36 percent developed or modified pipe-hanger hardware for their own use.

There's no mystery why this should be. In general, we all share the same needs, but specific requirements vary in ways that motivate a shift from mass production to mass customization. The question is how to do mass customization economically.

Von Hippel advances the notion of user innovation toolkits. The Apache Web server, with its modular architecture, is an example of such a toolkit. In the hands of skilled programmers, Apache can be, and often is, tailored to specific needs. When such customizations are shared, other users benefit. But so do Apache's developers, who, by observing what's done with the toolkit, can more intelligently evolve the core product.

Web-based software delivered as a service is, at its best, another kind of innovation toolkit. Users bring the data; developers wrangle the code; more useful innovation happens faster than it otherwise could.