What Does "Generative" Mean Anyway?

Today's Sunday New York Times included a column of mine in the business section which tried to think about the App Store's tremendous rate of innovation over the past two years, and the clear benefit it has had for small developers. My general point was that, while many of us have worked under the assumption that open platforms tend to be more "generative," in some important ways, the closed architecture of the App Store and the iPhone OS has actually contributed to the generativity of the platform.

On Twitter this afternoon, I got into a bit of a debate with Dan Gillmor, whom I have admired for many years now, about my use of the term "generative." Twitter is a fantastic medium for sharing links to complicated ideas, and an atrocious medium for actually discussing complicated ideas, so after a few volleys with Dan I told him I wanted to retreat to the blog. So here we are.

Dan's original tweet read as follows:

it's a flat contradiction to call a system generative, as @stevenbjohnson does, if it has a corporate choke point 

Now, both Dan and I are using Jonathan Zittrain, and his excellent book, The Future Of The Internet and How To Stop It, as a reference point for this concept of "generative." Zittrain's book is indeed about the power of open platforms that lack "corporate choke points," and generativity is inarguably crucial to his argument. But I don't think the two are synonymous. Or at least I don't think they should be synonymous.

In the opening chapter, Zittrain introduces the concept of generative platforms with a discussion of the Apple II: "[it] was a quintessentially generative technology. It was a platform. It invited people to tinker with it. Hobbyists wrote programs. Businesses began to plan on selling software. Jobs (and Apple) had no clue how the machine would be used. They had their hunches, but, fortunately for them, nothing constrained the PC to the hunches of the founders."

With the exception of that last clause (more on that later) all of these statements are true of the iPhone OS. (It is not true of the iPhone or iPad hardware, to be sure, as Cory Doctorow has emphatically pointed out.) As I argued in the column, the iPhone platform has seen more tinkering and new uses, generated by businesses and hobbyists alike, than any other two-year-old computing platform in history. (Including the Web, I would argue.)

In my mind, a "generative" platform translates to something like this: a platform that is constantly being re-invented in surprising new ways by a diverse group of creators, where individuals, hobbyists, small startups, and amateurs compete on a level playing field with large incumbents.

In other words, generativity is the result we are aiming for. An open platform is the tool we use to achieve that result. That's the central argument of Zittrain's book: if we want generativity in our media and software, then open platforms are the best way to achieve that goal. But if open platforms are part of the definition of generativity, then it suddenly starts to sound a little circular: the best way to achieve a generative open platform is to build a generative open platform.

What I was trying to argue in the Times piece is that the goal we're all championing—rapid, emergent innovation with small guys competing with big software companies—is happening on the iPhone platform, and it's happening in part because of the way that platform is closed, not open.

Now, that last clause: "nothing constrained the PC to the hunches of the founders." It is true that there is a very real constraint built into the App Store approval process, and in the column I specifically singled that out as the biggest problem with Apple's current system. But in practice, I would wager 99.9% of all new ideas get approved for the App Store, and so it is really more of a threat of a chokehold thus far. (Though I would still like to see a side door, as I wrote.) But of course these things all exist on a sliding scale. There are entirely open platforms that are entirely non-generative because no one develops for them. Apple's argument is that some restraints, implemented in the right places and judiciously used, can lead to more innovation in the long run. And right now, given the track record of the past 22 months, I think that's an argument that we have to take seriously.

Comments

[osandi]

i think of constraints as the pressure needed to push an idea to the point of function or practice.

my challenge for instance has been with open source hardware controllers that are designed to allow any user create the parameters for music software like ableton live.

with over 80+ buttons and knobs and very little experience with the software (yet more w/ protools), it's actually been more limiting (no to my ideas but how to implement them) than if i was restricting sections for a specific use.

great design has it's laws, like those of nature which if anything challenge the notion of 'can this happen here?'.

very very very insightful analysis people.

i love thinking on this level.

osandi

twitter.com/osxyz

[Robert Goldman]

I own both an ipod touch and a Google Android phone, and I think it would be instructive to compare the experiences on the two platforms.

While I am a strong proponent of open source tools, use them every day in my work, and do what I can to contribute to them, the Google Android experience is a very strong argument against unfettered openness.

The Android Marketplace is mostly a trash heap. It is well-nigh impossible to navigate, and the software one downloads seems as likely to supply an unpleasant disease as a useful service. The only tool that we are given to winnow through this enormous pile is a set of reviews by other users.

These reviews seem almost completely useless. They are summarized in a one to five star ranking that is not calibrated to any standard of quality. There is no way to tell who the raters are, or whether any individual rater is to be trusted.

For a text edit application, the top comment is "cant[sic] display chinese" (2 stars). Another is "PLEASE IMPLEMENT a SEARCH FEATURE...absolutely essential..." (5 stars). Are we to expect the application to get six of a possible five stars should this "essential" feature be added. There's a thoughtful and mostly positive review of the interface (3 stars). "Not wat [sic] I exscpected [sic] uninstalled." (1 star). Another app: "Force closes [crashes] way too much..." (3 stars). Software that crashes frequently, should never be three stars.

Wading through this for a while makes me understand why, after Yeltsin, the Russians flocked to Putin. It makes me want someone to impose some kind of standard. This is what happens when openness is not coupled with any standard-setting mechanism (such as, for example, the linux implementation assemblers like Ubuntu, SuSE, etc.).

I encourage you to have a look at Android; it might help bring a new perspective to this discussion, if only because it shows openness is not enough.

[Carla Casilli]

It appears that you, Mr Zittrain, and Mr Gillmor are arguing politely and intelligently about Apple's system—what is essentially a walled garden. Generativity, emergence, lock-in: these are large and important issues so I am grateful for the discussions occurring around them now. I only hope that they're landing in the right ears.

In the midst of this ongoing dialogue, let's keep in mind the observer effect, or the Heisenberg indeterminacy effect, or the observer-expectancy effect (it goes by any number of names), which suggests that what is watched is directly affected by the watching. Whether this effect will work toward increasing application creativity or decreasing consumer options has yet to be seen. Apple is certainly implementing a type of communication control. And this leads me to ask the ultimate question with regards to Apple's vetting process: who will watch the watchers?

[howardweaver]

I’ve gotten into hot water with Dan Gillmor, Dave Winer and others I admire (on Twitter and elsewhere), for referring to “the religion of openness.” Still, that remains the most illuminating metaphor I’ve come across.

In the temple, open is simply good. To whatever degree something is less than open, it is bad. Results do not matter: it’s ideology.

Me, I’m an agnostic. I’m not interested in Apple’s sins, or Google’s sainthood, or vice versa. What works? What yields the best results?

Isn’t that the scientific method?

[kathy gill]

Steve, I think the more troublesome point is the possibility that publishers might move content off the web and onto iPhone applications. Certainly, they are unlikely to do so in the short run, since the universe of people who read news sites on a computer+browser is much larger than the universe that does so on an iPhone, iPad or iPodTouch -- whether on browser or app. But that threat is, I believe, a real one.

It is that potential balkanization of content -- off the web and onto iPad apps -- that causes my angst. Look at the WSJ iPad app: no highlight/copy, no visible URLs to snag. A very tightly controlled one-way communication environment: old school, in other words.

Regarding user experience, however, I think the constraints that Apple has fashioned around iPhone apps are comparable to the constraints that made Mac computers "just work" when the PC world was an anarchy of mismatched software and hardware. The mainstream consumer benefits -- the lunatic fringe (edge case geeks) and early adopters, not so much.

An anecdote about controlling the store: I was watching a Harvard Business School video this weekend - one on the continuing gender gap in salaries. What ad was served up before the video? One for Trojan condoms! I don't think that Harvard would have been very happy had they known this -- does that make them "bad" should they decide to exert control over the user experience associated with their videos? I think not, at least in this case. And that's how I think about the iTunes store.

Disclaimer: I own AAPL stock. And my S.O. works for MSFT.

[Rikardlinde]

Clearly there's a lot of iPhone innovation going on and the amount of apps is a good sign but I'd like to know more about two things.

1. What kinds of apps are being built for iPhone compared to the web? There are, for example, thousands of innovative games but only 2 web browsers. Maybe not a fair comparison but I'd like to know more about the distribution of different kinds of iPhone apps.
2. What is Apples message to developers who want to extend the functionality of another app?

-Rikard

[twitter.com/mitensampat]

Steve, your definition of generative is more befitting the underlying meme of this discussion.

A platform that creates a level playing field for participants of all sizes while delivering a great customer experience (or value of some form) is the key ingredient for it to be generative.

the success of a platform (whether the iphone OS, or the Internet) depends largely on customer adoption ...and not on a set of legalese that protects one side over another. customer adoption directly depends on the customer experience which includes many things outside of the developer agreements.

Apple is doing a great job of moderating all the above thus far.

-Miten

[Jeanot van Belkom]

Steve, when you state that generative platforms are the result we are aiming for, I have the feeling you are mixing up two classes of arguments. Yes, the result also matters, but there's always the underlying fundamentals. Take the constitution: perhaps in every practical sense, a country may enjoy complete freedom of speech. But that practical situation is no substitution for the fundamental right laid down in a constitution. A monopoly may also result in a de facto generative platform, but it's still not desirable to have one. It's all about control - no matter how good your behaviour, it never makes up for a fundamental flaw.
- Jeanot

[Game Controllers]

Very informative article!!
i was bother to these What Does "Generative" Mean Anyway?
i think the best way to achieve a generative open platform is to build a generative open platform.

Nice article ..

[jordan 11]

Mishaps are like knives that either serve us or cut us as we grasp them by the handle or blade. Do you understand?

[tera gold]

Generative led to our goal. An open platform, is a tool we use to achieve this result. This is Zittrain's book of the central argument: if we want in our media and software, generative, and then turn on the platform is the best way to achieve this goal. However, if the generation of open platform part of the definition, then it suddenly started to sound a bit circular: the best way to achieve the formation of an open platform is an open platform to generate.Yeah,that is right.

[seks izle]

I own both an ipod touch and a Google Android phone, and I think it would be instructive to compare the experiences on the two platforms.

[jordan 7]

When talk about field,the first sight I think that is the own of somebody,but now I haven’t thought that,because I read your post,that new opinion there,thanks,learn the knowledge from you lots.