The growth of the open source software development movement is held up as one of the great successes of a networked world – leaving source code open is associated with global-scale participation in software development and open-source products that are now central to the technology industry. This has in turn inspired calls for opening up other “closed” processes – government, education, knowledge (like Wikipedia). There’s now talk of a global “open everything” movement.
But as Steven Weber explains, there are some specific elements to open-source development. First, that open code is a way of providing easily modifiable basic tools that can be customized to solve a whole set of different problems. This is one key to the success of open source – it’s the utility of the source code that’s available, and ability to modify it. So my friend who is working on a totally bespoke database can draw elements of source code from other databases built by others, even if those other products have little to do with what he’s making. Weber’s second element is that open-source is based on principles and values rather than efficiency.
Given these key elements, can we expect to produce “open everything’? Under what circumstances does an open-source model translate outside of software? To investigate this I’ve started watching the nascent movement towards open hardware development. Of course, hardware is a physical product with manufacturing costs. But if we think the design and production process, there are some clear opportunities to create an open source production ecology.
First, hardware designs are not material objects. They are, like software, intellectual products. Currently, most hardware production is based on patented designs. But hardware hackers (or hobbyists) can upload, view and download designs at OpenCores, which also allows would-be manufacturers to produce prototypes of their chip designs. Second, the realms of software and hardware are converging. The cost of developing software-controlled chipsets is dropping, with the major cost now being the software development itself.
The larger issue is how to grow an open source development and production ecology. In software development, one aspect of this ecology is the licensing framework, which identifies free software and makes using source code conditional on releasing any subsequent source code. How could this happen in the hardware world? How would a prospective hardware (re)designer know that the amazing mobile widget she/he was holding had an open design?
The solution, according to a nascent coalition called the Open Hardware and Design Alliance (OHANDA – watch this space) would be to develop a trademark sticker, to identify a piece of open hardware. The sticker would include a registration key, pointing to a design held in a repository somewhere. Then that design could be reused.
This potential intervention raises some interesting questions about “open everything.” How do open ecosystems grow? How modular do the “open” elements have to be? (it would be obviously more valuable to have a few, easy-to-use open hardware models than one design that’s difficult to reuse). And finally, what are the defining values of openness? OHANDA may provide some important lessons.