Zittrain gets me thinking about generativity and aggregation

My colleague Jonathan Zittrain has an important new paper published in the latest issue of the Harvard Law review. While it’s a paper for lawyers and legal scholars, it’s a critical read for anyone who’s interested in rigorous thinking about the Internet, its potential and threats to its future.

JZ believes that the most interesting feature of the Internet we know and love is “generativity” – the ability for “unrelated and uncredited audiences to build and distribute code and content through the Internet to its tens of millions of attached personal computers”. Technologies are generative when they are:
– Leveragable (they allow you to accomplish tasks that couldn’t otherwise be accomplished)
– Adaptable (useful for multiple purposes)
– Easy to master
– Accessible to a broad audience.
Defined by these criteria, an interconnected network of multipurpose devices which can be repurposed by third-party code is one of the most generative inventions ever created. (Written language gives it a run for the money.)

Because this is a law review article, it’s insufficient to celebrate the generative capacity of the Internet – instead, a threat must be confronted, solutions proposed and dragons slain. JZ sees a future where the generativity of the Internet comes under attack, from governments who try to protect users and users who try to protect themselves. The threat is generativity itself – because computers are programmable multifunction devices, and because the Internet can distribute code far and wide, it’s possible for people to create viruses and other “badware” that makes computers slower, destroys data or otherwise makes users unhappy.

(In true Berkman spirit, JZ’s response to badware isn’t just to theorize – he’s part of the team working on StopBadware.org, a project which is documenting software that attempts to hijack your machine in one fashion or another.)

Once enough computer users get sick of virii and spyware on their machines, they’ll do one of two things, JZ believes: demand law that restricts the ability of programmers to create certain types of software, or gravitate to less generative devices – special purpose appliances that perform a limited number of functions, rather than any abitrary, new function. (An example might be a Tivo, which is a special function device more powerful than an unaugmented VCR, but much less powerful than the powerful and general purpose MythTV.) While Grandma might be happy with an easy to use, uncrashable email-only device to communicate with her grandkids, the adoption of devices like this means that Grandma can’t make free Skype phonecalls to you, because her device can’t run that piece of new software.

JZ offers some interesting solutions to this problem – which you, like me, may want to challenge as a strawman. (After all, virii have been around for a long time, but you’re reading this post on a personal computer, not on a WebTV.) One involves challenging the importance of network neutrality above all else – maybe we want ISPs filtering our email and web traffic for virii and spyware if most users aren’t smart enough to filter at the endpoint, their own PCs. Others involve a system of virtual machines – a “green” machine that runs only carefully approved code and is highly stable, and a “red” machine where you can run experimental code, and run the strong chance of crashing your machine. The most interesting idea, in my opinion, is a distributed warning system that monitors computer behavior (should you choose to participate) and reports on software that appears to make computers less stable and users less happy.

Reading this paper, you can come away with the conclusion that generativity is mostly about the abililty of computers to execute arbitrary code. Talking with JZ the past two days, it’s increasingly clear to me that his thinking about generativity is much larger, and includes the types of generativity I’m most interested in – the ability of the read/write Internet to allow users to create and share content.

While I’m not planning on moving to an “information appliance” to replace my PC any time soon, I’m less worried about a possible move to less-generative devices than JZ is. Much of the great creativity we’ve seen on the web has happened on the server side, not the client side. I think some users will move to less flexible devices than conventional PCs, but will contribute to the diversity of the Internet by creating original text, pictures, audio and video – so long as they’re able to create this content and share it online, I’m less worried about whether the devices they use to edit and upload it are arbitrarily programmable or not.

What I like about JZ’s generativity concept – whether or not I subscribe to the problems or solutions he posits in this particular paper – is that it gets closer to what’s great about the Internet than other thinkers have come. He challenges much of the accepted wisdom of Internet commentators – not just a strong network neutrality argument, but the importance of open source. What’s important to him is the capacity for a programmer somewhere to write a piece of code for a system – open source or closed – and have other users adopt and use that technology. JZ suggests we judge technologies not on whether they’re open or closed source, or whether they’re network neutral in the strict sense of the term, but whether they contribute to the overall generativity of the network.

It’s a provocative line of thinking… and one, I think, that can serve as an organizing philosophy for much of the work we do at Berkman. The work Rebecca and I do on Global Voices celebrates the generative capacity of content creators around the world, and attempts to encourage further generativity by amplifying the voices of people who’ve chosen to use these tools to express themselves. Open Net Initiative documents barriers to the Internet that prevent people from encountering content and from expressing themselves. The folks who work on network neutrality are concerned with preventing changes to the networks we rely on that make it harder for people to create new content and code.

What’s great about generativity as a central principle is that it’s a positive idea we can embrace, rather than fighting battles against multiple foes real and imagined. It’s easy to rally people against censorship and net filtering, harder to rally then against changes to network neutrality rules, and very hard to rally folks in support of the ability to speak and code anonymously. It may be more productive to rally people in favor of the power of the Internet to enable creativity.

Talking this idea over with colleagues the past two days, I find myself wondering whether some of the lessons we’ve learned from Global Voices have applicability to other work organized around generativity. One of the most powerful lessons we’ve learned so far is about the power of aggregation. If you accept the idea – as I do – that the Internet is changing radically and that an increasing number of people will create code and content, not just consume them, being able to collect and curate that content is an important function. Colleages like Jake Shapiro at Public Radio Exchange explicitly aggregate content – in their case, programming that public radio stations can purchase for use in their markets. Projects like StopBadware are moving towards this model, aggregating new software and flagging the bad stuff. If Berkman moves towards a model where generativity is an organizing idea, it’s possible that the work we’ve done at Global Voices, figuring out how to aggregate and filter voices, may become central to Berkman’s working methods…