Feedonomics

Warning: This post is more pedantic than usual. Yes, that is possible.  You can tell you are in trouble when I start talking about graduate school.

I went to a Tweetup at Joe Cascio’s house yesterday, and thinking about the people who were there has made me realize that I’ve been simplistic in the way I describe generational differences in the software world. My grad school professors often accused me of simplistic thinking, and kept telling me to surprise them. At first I thought they wanted me to pull some obscure citation out of my butt, but what they really wanted was for me to stop using off-the-shelf ideas to explain history. One way to get historians excited is to find a new form of periodization, and while driving back yesterday I came up with a more informative model for breaking the history of personal computer software into two distinct periods. I typically divide software people along traditional generational lines: Baby Boomers, Gen X and Gen Y. It may be more appropriate to separate developers based on whether they came of age in the time of InfoWorld or Slashdot. It doesn’t matter how old they are, as much as the first full business cycle they experienced. A friend of mine who used to manage radio stations once told me that people’s musical preferences are formed in high school, and tend to freeze at about age 18. A similar pattern exists for software people who either started in the Eighties era of desktop client software, or with dot com software in the Nineties.

The InfoWorld generation thinks of themselves as entrepreneurs, and have Dan Bricklin and Mitch Kapor as role models. Their foundational creation myth starts with Bricklin in a Harvard Business School classroom. Following business school dictates, they believe that they must  find an underserved niche, and try to build a company that delivers a product to fill that niche. The company is the basic unit of creation, and the first product is just the start of a long product line that will lead to an on-going, profitable business. They saw their first cycle come to a close with the death of Ashton-Tate and Lotus, and the establishment of Microsoft as the pre-eminent software company. This didn’t mean that their ideas about creating profitable companies were wrong. On the contrary, they believe that Gates won because he was better at following their model than anyone else.

The Slashdot generation think of themselves as coders, and have Linus Torvalds and Eric Raymond as role models. Their creation myth has Linus creating a variant of Unix as a hobby, and then giving it away. They believe that software should be created to fulfill a personal need first, and then given back to the community, which will obviously have similar needs. The InfoWorld generation often dismiss this thinking as “communist,” but Slashdotters are not anti-capitalists, they are anti-corporatists. Corporations are evil, and Microsoft is the most evil corporation of them all. By giving away their code, Slashdotters gain in stature, and eventually become a financial success as well, but without the trappings of a big company run by “suits” from the InfoWorld generation. They saw their cycle come to an end when the suits were proven to be frauds, and Google, a company that was run by coders like themselves, and a sworn enemy of evil, was left holding all the pieces. Ironically, or maybe appropriately, their cycle ended right after Slashdot went public, an event I had a role in.

So what can we do with this model of two generations? For one thing it helps explain the disconnect between VCs and software developers in the current Web 2.0 cycle. It isn’t about software being easier and cheaper to build today. Those are fallacies, as plenty of Web 2.0 companies are now discovering. They are based on Slashdotters valuing their own labor at zero cost, because they are doing it out of love for the code and their community. When InfoWorlders get involved with running these companies, they quickly discover that paying coders still costs real money, and making quality software still takes a long time. The real problem is that InfoWorlders may have access to financial resources, but they no longer have the coding chops that they need, so they have to hire Slashdotters who have a completely different world view. Try reading “Dreaming in Code” to see what happened when Mitch Kapor tried to emulate Linus Torvalds.

Can this theory of two generations be used to predict the future? Sorry, I’m a historian, not a consultant. There are plenty of consultants who think they have the answers already. Who knows? They may even be right. I was taught to wait until a cycle is over before trying to decide what the secrets of success were.