Web Services and Open Source at OSCON

I spend a lot of time with my head buried in code, and every time I pick my head up it feels like the future is closer than I thought. So I like coming to OSCON. A week of looking ahead leaves me more confident I won't get future shock anytime soon. OSCON, like all conferences, is aimed at corporations, the intangible entities that send humans as their proxies. But open source has its roots in individuals working outside the corporation for their community of programmers. Are the two cultures coming together, or colliding? And how will the "open source ideal" evolve, as the chief social act of programming changes from trading disks of source code to processing each others' data and mashing up web APIs?

I'm an open-source programmer who's lucky enough to be paid by a corporation. Between sessions this week I'm working on turning Slash's metamoderation into a plugin, making Slash more useful for other site admins. I'm a human first and employee second. And I'm concerned about how the community based around this software ideal of not welding shut the car's hood is going to hold together.

Markets aren't designed for goods with zero cost of reproduction, but because property is such a powerful tool for efficiency and prosperity, societies have been artificially constructing markets for creative works since even before the founders wrote up their support for "science and the useful arts." Often, markets in ideas work pretty well.

There have been three societal "bow shocks" in the collision between programming and capitalism. The first hit in 1976 when Bill Gates insisted that charging for software made sense. The second was in the late 90s when open source proved better than corporate hierarchy at certain types of development. And then there's the one that's about to hit now, when web services and interoperability concerns obviate open source licenses.

There's a growing understanding here that web services are big: that the laptops and desktops of the future will rely not on software goods that have been bought for those machines, but software services that run on a server a thousand miles away. Google calls its Ajax web services "the world's largest platform."

Yesterday, Tim O'Reilly hosted a stimulating all-day series of panels and talks on web services and "Web 2.0" generally. The most interesting part of the discussion was about tying web services together. Web mashups are hot. It's hard to look at a list of websites offering an API -- Google Maps, Yahoo Geocoding, eBay, craigslist, Flickr, YouTube -- and not start thinking about great ways to combine them. Interoperability plus programming creativity equals... well, something pretty neat, we're hoping.

But a web services API doesn't necessarily offer the freedom that might seem analogous to open source, which is why Tim is also putting out the call for an "open services" definition. Flickr offers its corporate API to some sites, and refuses to permit it to others. Zooomr was judged to be too much of a potential competitor, so Zooomr users don't get to copy the photos they've uploaded to Flickr. [Update: Sorta. Read that comment thread to see important context for Flickr's decision. To be clear, given that context, Tim thinks Flickr found a good answer, and I tend to agree.]

As Flickr says, and they have a very good point, "why should we burn bandwidth and CPU cycles sending stuff directly to [a potential competitor's] server?" That makes sense from a corporate point of view, but a user who's uploaded a thousand of their photos might be puzzled why it's no longer exactly "their" data. Is that a right that user should have, or not? I ran into Julian Cash, who vehemently argues that it is; he's started MoveMyData.org to try to build a client-side way for users to route around APIs, to suck down "their" data and maybe reupload it to other sites. No code yet, but he's looking for volunteers.

AttentionTrust goes even further, starting off its manifesto with "you own your attention and can store it wherever you wish." That's something I hadn't considered before but it has an interesting ring to it. They have a Firefox extension I haven't tried yet (does it work? post comments).

Interoperability is a concern even without the web. Yesterday morning, Danese Cooper got a half-hour to grill Bill Hilf, Microsoft's General Manager of Platform Strategy, on Microsoft's relationship with open-source. Some think that's the same relationship as the butcher to the hog, and Bill's job is to persuade them Microsoft has no such intentions.

Asked directly, in the context of embrace-extend-extinguish and web APIs that can be crushed at any time, "why should we trust Microsoft?", Bill's answer was to look at the company's actions: "consistent action, over time, in the right direction."

I sat down with him afterwords to probe into this a little more (with someone from Waggener Edstrom standing nearby). He has some examples of Microsoft working with open-source projects like JBoss and SugarCRM, but I asked for specifics of how we know Microsoft isn't going to try to kill more-directly competing projects like Mono or OpenOffice by eliminating interoperability, possibly with patents, at any random time in the future. The only real sign I got was the Covenant Not to Sue (over patents) that came with the OpenXML format earlier this year. That's a step in the right direction. I don't think it's a terribly big one.

I asked if we'd see more steps from Microsoft disavowing patents as weapons against open-source projects. Obviously that's a big risk for a company to take, but one that's probably necessary to convince skeptics Microsoft is friendlier than the butcher. While Bill couldn't make any promises, he affirmed the CNS was "not a one-off... and not just to placate people." I'll keep an eye out for more action in the right direction.

Exciting as the opportunities are for different projects' software working together, one thing's for sure: the remote sites that run their algorithms and store your data leapfrog open source licensing. The server a thousand miles away can run software with its hood welded shut, with no obligations to the open-source community that come along with the benefits. Today, while some companies are trying to build goodwill with that community, there is nothing like a GPL for web services. No one's discovered a legal foundation that would establish open services, openly shared web services, with the same kinds of rights that we insist on in open-source code. No one's even sure what "open services" might mean, indeed, there's no consensus that we even need such a thing.

Even the FSF is unable to decide how v3 of the GPL should read. And I'm not smart enough to know if the GPL is even the right tool for this. Maybe tacking clever licensing terms on top of copyright's restriction is a temporary hack whose time has passed (you know, like the RIAA). Maybe the next hack to build a community of software sharing and tinkering will have to be totally different.

I don't think I know the answers but maybe one of you does. If you have thoughts about the open-source community in the age of capitalism, please post them to this story. If you're at OSCON and want to chat about it, email me (or AIM 'jamiekzoo' if you catch me online). At the end of the week, I'll have more updates on what's happening here -- it's not all philosophy and futurism.