Open Source Hardware
This article first appeared in the January & February issue of Elektor.
Open source software has proved it’s worth its salt. Firmly rooted in two of humanity’s more profound qualities, curiosity and collaboration, open source is driven by a need to access, share, modify and improve the source code. Because open source allows people to build on each other’s solutions, it prevents duplication, cuts development costs, lowers entry barriers, and generally makes programming a lot more fun. With over...
Open source software has proved it’s worth its salt. Firmly rooted in two of humanity’s more profound qualities, curiosity and collaboration, open source is driven by a need to access, share, modify and improve the source code. Because open source allows people to build on each other’s solutions, it prevents duplication, cuts development costs, lowers entry barriers, and generally makes programming a lot more fun. With over half of the total number of mobile phones running on a Linux-based operating system it’s safe to say that open source software has decisively established its place in the world.
For all of its success it isn’t surprising to see the open source mindset ported to the hardware community and witness the emergence of Open Source Hardware (OSHW).
OSHW is hardware of which the design is made publicly available. The design may be shared, modified and (evidently) used to build hardware.
To make sharing and production more streamlined the OSHW community is working on standardization. They’re developing collaboration tools and design software. And inspired by open source software they’re publishing licenses which govern the rights and the restrictions of the user.
Operating at the outer reaches of OSHW is Sébastien Bourdeauducq, founder of the open source project Milkymist. He also organized several OSHW conferences, including the Exceptionally Hard & Soft Meeting which took place in Berlin December 28-30, 2012.
In a Skype chat Sébastien told me about opening up chip design and how the Milkymist code ended up in low Earth orbit.
Open Source
Tessel: How did you get involved in Open Source Hardware?
Sébastien: I've always loved to understand how things worked. I started doing electronics and programming when I was a kid. My first contact with actual ‘open source’ was in 1999 when I switched to Linux — but I had touched other people's code before, which was published in magazines, books and BBSs [Bulletin Board Systems] even though it wasn't labeled ‘open source’ or ‘free software’. I really liked Linux, after being frustrated with the restrictions of Windows that prevented you from touching the core of the operating system.
My first project that could be called ‘open source hardware’ was in 2002 — it was an infrared communication system for the TI89 calculator, with GNU Free Documentation License, a very simple thing. Then I went into wireless driver writing for FreeBSD (an open source UNIX-like operating system), which was more a mix of hardware and software, since I had to understand how the undocumented wireless hardware worked. But my biggest OSHW project today is Milkymist, which I started in 2007 (see inset). I wanted to open up chip design, something that few open source projects tackled and that's still the case today.
Innovate at the next level
Tessel: Why did you want to open up chip design?
Sébastien: One reason is philosophical — almost everyone's so called ‘open source’ electronics today is built on black box chips that they don't understand, and they don't try to. I found this barrier to understanding very frustrating.
Another reason is security: you can actually have hardware backdoors. (A backdoor is a manipulation of a chip’s firmware or hardware enabling an unauthorized party to gain access and disable or reprogram the chip. Backdoors can be installed on a chip during the production process. They’re considered a genuine threat, especially in military hardware.)
Finally, the last reason is being able to innovate at the next level. This is actually quite successful, but outside the common OSHW community. Milkymist code has found its way into major scientific projects like the International Space Station (ISS) and particle accelerators.
In 2009 NASA’s Jet Propulsion Laboratory (JPL) used the code of Milkymist SoC memory controller for its Space Communications and Navigation (SCaN) Program. On October 15, 2012 Sébastien received a letter saying:
Tessel: You also contribute to open hardware initiatives of CERN, the European Organization of Nuclear Research. Many research institutes are secretive about their work, why does CERN embrace the OSHW approach?
Sébastien: Well, there are many different personalities in research institutes. It's true that many are quite secretive — sometimes on purpose, sometimes because they don't care and then academic publishers put research behind pay walls to make money. I think the high energy physics community is a little more open than the others, maybe because they do fundamental research that has no immediate commercial or military applications.
The CERN open hardware initiative is also the work of a handful of employees who believed in OSHW and wanted to make their institution part of it. They have created an online Open Hardware Repository where electronics engineers can collaborate on open hardware designs used in experimental physics. And in 2011 they published the CERN Open Hardware License.
Exceptionally Hard & Soft Meeting
Tessel: Why are you organizing the Exceptionally Hard & Soft Meeting (EHSM)?
Sébastien: Many hacker and OSHW conferences, including the three I have co-organized before EHSM, lacked in some areas of content. Most of the hardware technology is extremely simple and carefully goes round expensive and difficult areas. Which is kind of frustrating for the same reasons I started the Milkymist project. EHSM wants to go pretty far, for example our keynote speaker is a 17 year old boy who built a nuclear fusion reactor.
EHSM is also trying to involve scientific institutions and research. We already have CERN. We'll probably have people from a quantum physics lab and perhaps even a hands-on experimental quantum physics workshop. We’re also talking to another institute in Berlin which does very advanced materials research.
Another change we want to make with regard to other conferences is that they are like 99% male. [Find streaming videos of the conference here (labeled EHSM) and torrents here]
Tessel: OSHW is different from open source software in that it will not only have to deal with copyright but also with patents. What are your thoughts on patents, do they harm OSHW?
Sébastien: Patents are so poorly implemented that they harm everything, including OSHW. One of the original ideas of the patent system was to encourage inventors to publish by giving them legal protection before the patent becomes public domain. Which is, in a way, an open source system because you make inventions public domain in the end, after the patent expires.
But with the patent system in place today, needless to say it failed miserably. It has just become some corrupt corporate battle ground. In particular when things that are not innovative at all get patented (see for example Apple's renowned patent on a rectangle with rounded corners). And only large corporations infringing on each other’s' patents have an advantage there — they agree not to sue each other. But newcomers are easy targets; as a consequence the patent system stifles innovation rather than promote it.
The future of OSHW
Tessel: If the world would go open source, what would be the benefits?
Sébastien: I would see it as a huge learning opportunity to start with. But economically, I don't know. It's not very clear to me whether open innovations and the fact that you can collaborate far more easily with more people outweigh the benefits you may get from keeping your design details secret. But I like experimenting, and that's also why I got into projects like Milkymist. The positive thing I can say so far is the OSHW approach has been successful in attracting talented people.
Tessel: What role will OSHW play in the future? Will it replace proprietary closed hardware?
Sébastien: I don't see it happening in the near future. The reality is that most hardware is manufactured by companies like Apple or Intel who do not care the slightest bit about OSHW and hold a lot of know-how that they are not ready to release. And most OSHW people -at least those who don't go to EHSM- do not care about creating industry-level hardware. Of course, hopefully it'll get better in the long-term but it will be hard and will need a lot of mindsets to change.
For all of its success it isn’t surprising to see the open source mindset ported to the hardware community and witness the emergence of Open Source Hardware (OSHW).
OSHW is hardware of which the design is made publicly available. The design may be shared, modified and (evidently) used to build hardware.
To make sharing and production more streamlined the OSHW community is working on standardization. They’re developing collaboration tools and design software. And inspired by open source software they’re publishing licenses which govern the rights and the restrictions of the user.
Operating at the outer reaches of OSHW is Sébastien Bourdeauducq, founder of the open source project Milkymist. He also organized several OSHW conferences, including the Exceptionally Hard & Soft Meeting which took place in Berlin December 28-30, 2012.
In a Skype chat Sébastien told me about opening up chip design and how the Milkymist code ended up in low Earth orbit.
Open Source
Tessel: How did you get involved in Open Source Hardware?
Sébastien: I've always loved to understand how things worked. I started doing electronics and programming when I was a kid. My first contact with actual ‘open source’ was in 1999 when I switched to Linux — but I had touched other people's code before, which was published in magazines, books and BBSs [Bulletin Board Systems] even though it wasn't labeled ‘open source’ or ‘free software’. I really liked Linux, after being frustrated with the restrictions of Windows that prevented you from touching the core of the operating system.
My first project that could be called ‘open source hardware’ was in 2002 — it was an infrared communication system for the TI89 calculator, with GNU Free Documentation License, a very simple thing. Then I went into wireless driver writing for FreeBSD (an open source UNIX-like operating system), which was more a mix of hardware and software, since I had to understand how the undocumented wireless hardware worked. But my biggest OSHW project today is Milkymist, which I started in 2007 (see inset). I wanted to open up chip design, something that few open source projects tackled and that's still the case today.
Milkymist
The Milkymist One is a stand-alone video synthesizer used for rendering live video effects at music performances. Both on the hardware and software level the Milkymist is open source where possible.
Remarkably, Milkymist has an FPGA-based System-on-Chip (SoC) that got custom developed and comes with a free Hardware Description Language (HDL) source code complement. Most components of the SoC, except the LatticeMico32 CPU core, are placed under the GNU General Public License. The SoC programming and functionalities are described in some detail in Elektor USA, June 2012.
The video rendering software Flickernoise is heavily inspired by Milkdrop, an open source plugin for the music player Winamp that generates visualizations. As open source projects go, many people have contributed to Milkymist.
Innovate at the next level
Tessel: Why did you want to open up chip design?
Sébastien: One reason is philosophical — almost everyone's so called ‘open source’ electronics today is built on black box chips that they don't understand, and they don't try to. I found this barrier to understanding very frustrating.
Another reason is security: you can actually have hardware backdoors. (A backdoor is a manipulation of a chip’s firmware or hardware enabling an unauthorized party to gain access and disable or reprogram the chip. Backdoors can be installed on a chip during the production process. They’re considered a genuine threat, especially in military hardware.)
Finally, the last reason is being able to innovate at the next level. This is actually quite successful, but outside the common OSHW community. Milkymist code has found its way into major scientific projects like the International Space Station (ISS) and particle accelerators.
In 2009 NASA’s Jet Propulsion Laboratory (JPL) used the code of Milkymist SoC memory controller for its Space Communications and Navigation (SCaN) Program. On October 15, 2012 Sébastien received a letter saying:
Dear Sébastien,
I’d like to take this opportunity to thank you for your open source work which helped our development of the JPL Software Defined Radio that is part of the SCaN Testbed, which has now been installed on the International Space Station. It has been undergoing initial checkout and commissioning for a few months now, including the use of the dynamic ram controller that Greg Taylor built starting with your work. I think it’s safe to say that many millions of bits have passed through the code you wrote in space by now.
Thanks again,
James P. Lux, Co-Principal Investigator, SCaN Testbed
Tessel: You also contribute to open hardware initiatives of CERN, the European Organization of Nuclear Research. Many research institutes are secretive about their work, why does CERN embrace the OSHW approach?
Sébastien: Well, there are many different personalities in research institutes. It's true that many are quite secretive — sometimes on purpose, sometimes because they don't care and then academic publishers put research behind pay walls to make money. I think the high energy physics community is a little more open than the others, maybe because they do fundamental research that has no immediate commercial or military applications.
The CERN open hardware initiative is also the work of a handful of employees who believed in OSHW and wanted to make their institution part of it. They have created an online Open Hardware Repository where electronics engineers can collaborate on open hardware designs used in experimental physics. And in 2011 they published the CERN Open Hardware License.
Exceptionally Hard & Soft Meeting
Tessel: Why are you organizing the Exceptionally Hard & Soft Meeting (EHSM)?
Sébastien: Many hacker and OSHW conferences, including the three I have co-organized before EHSM, lacked in some areas of content. Most of the hardware technology is extremely simple and carefully goes round expensive and difficult areas. Which is kind of frustrating for the same reasons I started the Milkymist project. EHSM wants to go pretty far, for example our keynote speaker is a 17 year old boy who built a nuclear fusion reactor.
EHSM is also trying to involve scientific institutions and research. We already have CERN. We'll probably have people from a quantum physics lab and perhaps even a hands-on experimental quantum physics workshop. We’re also talking to another institute in Berlin which does very advanced materials research.
Another change we want to make with regard to other conferences is that they are like 99% male. [Find streaming videos of the conference here (labeled EHSM) and torrents here]
Tessel: OSHW is different from open source software in that it will not only have to deal with copyright but also with patents. What are your thoughts on patents, do they harm OSHW?
Sébastien: Patents are so poorly implemented that they harm everything, including OSHW. One of the original ideas of the patent system was to encourage inventors to publish by giving them legal protection before the patent becomes public domain. Which is, in a way, an open source system because you make inventions public domain in the end, after the patent expires.
But with the patent system in place today, needless to say it failed miserably. It has just become some corrupt corporate battle ground. In particular when things that are not innovative at all get patented (see for example Apple's renowned patent on a rectangle with rounded corners). And only large corporations infringing on each other’s' patents have an advantage there — they agree not to sue each other. But newcomers are easy targets; as a consequence the patent system stifles innovation rather than promote it.
The future of OSHW
Tessel: If the world would go open source, what would be the benefits?
Sébastien: I would see it as a huge learning opportunity to start with. But economically, I don't know. It's not very clear to me whether open innovations and the fact that you can collaborate far more easily with more people outweigh the benefits you may get from keeping your design details secret. But I like experimenting, and that's also why I got into projects like Milkymist. The positive thing I can say so far is the OSHW approach has been successful in attracting talented people.
Tessel: What role will OSHW play in the future? Will it replace proprietary closed hardware?
Sébastien: I don't see it happening in the near future. The reality is that most hardware is manufactured by companies like Apple or Intel who do not care the slightest bit about OSHW and hold a lot of know-how that they are not ready to release. And most OSHW people -at least those who don't go to EHSM- do not care about creating industry-level hardware. Of course, hopefully it'll get better in the long-term but it will be hard and will need a lot of mindsets to change.
