Open infrastructure and the common good

We need to talk a little about something that makes Americans in particular, well, uncomfortable: a common good that can’t be adequately addressed by the marketplace.

At this very moment you’re using a magnificent outcome of this kind of “common good” approach that I’m talking about—the Internet. Yes, yes, I’ll pause for you to crack an Al Gore joke here, but let’s not miss the point. The Internet exists the way it does because no private or state actor owns it, right? The reason no private or state actor owns it is because of explicit decisions made by both its creators and funders to treat it as a common good. From TCP/IP up to higher-level protocols like HTTP and electronic mail, no company or government agency has the power to declare “from this point in time forward, things using this protocol will be different.”

Those protocols are open infrastructure. Sometimes they have nominal owners but control has been relinquished to a standards body; sometimes they’re true public domain. Businesses can build on them, governments can try to spy on them, and of course vice-versa—but they’re public roads, not private ones. Everybody can use whatever web browser they want or email client they want or MP3 player they want. People can (and do) build businesses on top of those protocols, just like businesses in the physical world are built on top of physical infrastructure that those businesses only pay for indirectly.

So. About Twitter.

Twitter isn’t the first proprietary protocol to take off; instant messengers are the most notable previous ones in my mind. But nothing else has become as entwined into the fabric of the mainstream so quickly. Twitter is being treated by the world at large as if it were in the same class as email and HTTP and IM. It’s on billboards, on fast food drink cups, in crawl lines across cable news channels. We all know it’s one company’s product, but it’s not the way we use it.

Suppose CERN had spun off a company headed by Tim Berners-Lee around, say, 1992. In this alternate universe, HTTP works the same way it does now—mostly. The difference is that WebCo™ authenticates all the browsers and acts as the CDN. If it’s done right it may even be a more efficient experience than what we have in the real world! But maybe WebCo™ decides, one day, that they need to ensure a “more consistent experience” for web browsers. Poof! No more ad blockers. (No ads without a cut to WebCo, either.) No plugins that reformatted web sites except in specific ways. No “read it later” services like Instapaper.

WebCo can enforce this because, well, they own HTTP. It’s not a common good in their alternate universe. It’s a private good. WebCo’s HTTP may be more awesome in many ways than the HTTP we really have, but when it comes to infrastructure protocol, a terrible open standard is better than a terrific closed one. The problem that Twitter is inadvertently demonstrating to us all is that they are to their own service what alternate-universe WebCo is to HTTP. recognized this issue very early on, but misdiagnosed the problem. People who aren’t nerds don’t care whether the software stack is open source, and would collapse if it was ever successful anyway. If you have a service that costs thousands of dollars a month to run (if not millions), your problem isn’t your software license, your problem is paying for your damn servers.

So what about Dalton Caldwell’s project? He’s trying to “fix” Twitter’s business model by moving it from advertiser-supported to user-supported, so it should never have the problem that wishes it would have. Yet as much as I’m with Ben Brooks in the “let me pay for it” camp (and yes, I’m an supporter, there as “chipotlecoyote”), I wonder if this is the wrong problem to be tackling.

If we think Twitter the service has long-term value as a fundamental Internet service—and I do—then Twitter the company needs to be disrupted. That’s not going to happen by building a clone with a more liberal software license or a better business model. It’s only going to happen if such a service becomes Internet infrastructure, no matter how many business models that breaks along the way.

Frankly, in today’s climate I’m not sure what company is going to step up to the plate to do that. In the long run, open infrastructure generates more revenue for more players than proprietary infrastructure does—but you have to believe creating a common good is more important than owning it.