Modularity, Openness, and Glass-Box Enclosure

Though I wasn’t able to attend this year’s Solid conference, I received messages from a number of colleagues there about one talk in particular, Kevin Czinger’s talk on Dematerializing Auto Manufacture. In that talk, Kevin compares the work he’s done at Divergent Microfactories with Arduino, referring to their system at one point as “Carduino.”

The talk is worth watching. In it, Czinger discusses the life cycle analysis of car manufacturing, and points out that the greenhouse gas emissions of electric vehicles over the course of their manufacture, use, and disposal is actually higher than the emissions for traditional internal combustion engine cars (I’m paraphrasing here; check his talk for the facts). He mentions that in the next 35 years, the world will likely produce four billion more cars, and that if we do not make significant changes now, the impact of that on the environment will be devastating.

He then goes on to talk about the work Divergent has done to develop a modular chassis system to reduce significantly the environmental impact of manufacture, if not use, of the electric vehicles they will produce. Discussing the modular system, he refers to it as “carduino” because it “hides the complexity and is simple to use.”

There are a number of points on which I agree with Kevin. I applaud Divergent’s work on life cycle analysis of their product, and I think the node system is very clever. I was struck by his forecast that we’ll see 6 billion cars in the next 35 years.  Is that inevitable, I wonder, or is that also a possible point of intervention?  Rather than accepting it, could we look for ways of changing the transportation industry itself?  But that is a moot point, since Divergent has decided to manufacture cars. For them, 6 billion cars is an inevitability.

I’m flattered by the comparison to Arduino, but I think Kevin is missing a key element of what I consider to be Arduino’s philosophy. It’s not just about simplicity and modularity, but also transparency. Though he mentions that they set out to build a modular system to simplify the design of cars, he doesn’t mention who will have access to that modular system. Nowhere in his speech does he mention sharing the knowledge of how they build their cars beyond Divergent, though he does invite listeners to join them.

I don’t think Arduino’s impact came only  from the fact that we made things modular and hid the complexity. Those are important features, sure, but we weren’t the only ones do to that, nor the first.  I think that sharing the plans for Arduino openly and producing plenty of examples early on had as much to do with its impact as the modularity and simplicity.

From the beginning of Arduino, numerous collaborators and critics have tinkered with, complained about, praised and improved upon the system. They just worked not just with the API, but also on the code behind that API (which has its origins in Programma 2003, Wiring, and Pascal Stang’s avrlib, among others), and on the circuit design and on the code and design of the IDE (which was built upon the foundations of the Processing IDE, of course). They are some of our most enthusiastic supporters and also our harshest critics.

We’ve also seen lots of people take their second step in learning about microcontrollers by diving into the details of that underlying code and customizing it. This is what I call “glass box enclosure” as opposed to “black box enclosure” — the box is still nicely packaged, but the complexity is not hidden, it’s just wrapped up so you can ignore the details when you want to. You can look inside the package if you want, and learn from it.

In taking this approach, we’re participating in a strong intellectual heritage, both of the open source movement starting with Richard Stallman and of the informal sharing of knowledge that often happens as colleagues move from one learning institution to another.

“Democratization” of a technology means that you allow others to get their hands on it and to participate in its development. It’s a messy process, and you open yourself up to lots of criticism in the process. You  also enable your competitors. Many would say it’s not good business sense. But if you’re lucky, your community and your competitors recognize that competition can be fair, that it can be collaborative, and that it can serve a greater good. No one person or one company can have all the good ideas for an industry. There are derivatives of Arduino which we do not make, but which are important and positive contributions to the ecosystem of which we are a part.  That’s what I think of when I think about democratization of a technology. 

I’m wholeheartedly in favor of companies who want to do this, and I’m hopeful that Divergent will democratize their industry. I think the ideals Kevin expressed in his talk about environmental consideration and systems thinking are great. Those ideals will spread if Divergent enable more people to practice them. They will create some fierce competition as well. Although Kevin didn’t explicitly say that Divergent’s system was open in his talk, he also didn’t say anything that precludes it, so I’m hopeful. I look forward to more concrete statements from Divergent in the near future. The publication of the plans for their modular system would be a great next step, so that everyone can democratize the car industry.

Knowledge sharing doesn’t have to take just one approach, however. Just as I’m hopeful that Divergent will make an open platform to support a thousand small car companies (similar to Local Motors, perhaps?) I’m equally enthusiastic about Tesla’s announcements about their patents. Although I’m an advocate and practitioner of open source hardware, I recognize that it’s got its limitations, particularly for larger scale enterprises. Tesla’s approach is an interesting back door to openness. Since patent filings are accessible through a country’s patent office, the method and system behind any patent can be obtained by anyone who wants it. By stating that they will not enforce licensing of their patents, Tesla’s saying “We explained how we did it in our patent filing, go read it and give it a go yourself. We won’t stop you.” Duplicating or building on that is not as simple as downloading an open schematic or source code, but it’s a step toward a more open culture in big enterprise.

To address the kinds of systemic problems Kevin Czinger referred to in his talk will take more than one company, one industry, or one nation. It will require lots of interlinked efforts and intensive interoperability of the things we invent. This can’t be done if each company’s bottom line is the only consideration. Cooperation is risky, but I believe it’s our most promising way forward.