The Linker is a blog from Christopher White and Elecia White, hosts and creators of the weekly embedded.fm podcast. The Linker discusses embedded software and hardware engineering, education, and entrepreneurship. The goal is to dive deeper into the tangential topics brought up during the show.
This post was written by Christopher White.
To hardware engineers, this must seem like a time of wonders.
As a software developer, I don’t have a lot of exposure to Open Source Hardware (OSHW). I’ve had a lot more to the Free and Open Source Software (FOSS) movement: I distinctly remember sitting in my dorm room in 1992, installing an early Linux distribution to my (then) cutting-edge 486DX2 PC’s second-hand 80MB hard drive (purchased for $120!). It was an incredible thing, moving from Windows 3.1 to what seemed like the future, and it was free (as in freedom and beer). My brother would log into it remotely from graduate school and make the floppy drive light blink just to annoy me.
That was the start of the software time of wonders. Experimentation became possible as prices dropped. Because of FOSS, we got good and free tools like Linux, the GNU toolchain/binutils, and FreeBSD.
To hardware engineers, this must seem like a similar time. No longer do you need a multi-thousand dollar license to use CAD software, because you can do quite well with KiCad. No longer does creating a board imply a huge financial investment or even a long lead-time.
In episode 92 we spoke to Laen Neal, founder of OSHPark. OSHPark started with the idea of aggregating small-run PCBs into a larger panel, sharing the setup costs over a number of projects [2:27]. Laen was inspired to help out a few hobbyists at his local hardware meet-up (DorkBotPDX). Like many electronics services that started small and local, OSHPark has become invaluable to a much wider market; its volume may soon justify the company becoming a fab in its own right [41:10].
At the risk of perhaps analogizing two movements that don’t have quite as much in common as their names suggest, I wondered out loud on the show [9:55] about where the OSHW movement was in its development compared to a similar time in FOSS’s history. Does today correspond to the late 90’s, when Linux was becoming a threat to mainstream operating systems (at least in the minds of the tech punditocracy)? Or is it even earlier, when the potential of Free Software as a force to change industry was still only a glimmer in the eye of ardent supporters?
Taking this analogy to the edge of the cliff, I am inclined to wonder what is the Linux of OSHW? Who is the Richard Stallman or the Linus Torvalds?
While similar big names may exist in OSHW (Arduino is certainly a good candidate for the “Linux” analogue), I’m not sure this movement needs them in the same way. When FOSS and Linux got their starts, they were competing for mindshare vs. monoliths: big companies like Microsoft, Sun, and IBM. Personalities and advocacy were required simply because the concept of “open” was so new. The competition for OSHW is far more distributed and varied, and it can build on the foundations and awareness the FOSS built.
Still, single products or services really do not define either movement. Both engender tools that can be employed by anyone to create. In software, the GNU toolset and the operating systems they run on can be used to build just about any code that can be put on a computer or a device. The compilers, the libraries and frameworks, and the infrastructure all are natural outgrowths of the philosophy. OSHW has inspired an ecosystem that parallels what the FOSS movement has produced, from design software (KiCad) and platforms (BeagleBone Black, Arduino) to manufacturing (3D printers, desktop CNC mills, OSHPark). These are all tools and services that used to be prohibitively expensive for the average person. Just look at what has happened to the price of development kits in the last decade!
FOSS allowed software engineers a way to experiment. The only downside to trying out an idea was time (which as a college student, I had). Mistakes were easy and fixing bugs was an educational opportunity. Historically, hardware has been less forgiving of mistakes, with high cost and long lead times. However today’s hardware improvements and new manufacturing methods allow for cheaper mistakes. This can only encourage exploration and experimentation. Growing marketplaces like Tindie, Sparkfun, Adafruit, and Pololu bear this out.
My parallel does break down in a few areas, though. While copyleft licenses like the GPL kept (for better or for worse, depending on your view) open source software from being easily mixed with or subverted into closed source in many cases, it seems a bit harder to apply that to hardware. One can certainly release schematics and designs under GPL or Creative Commons Attribution-ShareAlike (which would require derivative designs to also be open), but discovery of violations and enforcement might be even more challenging than with software. It would be interesting to compare the ratio of software projects released under permissive licenses vs. that of hardware projects. My guess would be that OSHW projects are more permissive, partly because of the influence of the Maker community.
Producing a piece of hardware also requires a mix of the digital and physical realms, which complicates it further. A truly open piece of hardware is elusive, as greater diversity of open MCUs and wireless chipsets is needed. Some of the tools associated with OSHW are also a mix of open and closed (that 3D printer on your bench might have open hardware, but closed software, for instance).
It’s encouraging to hear Laen say that OSHW is trying to learn from the experience of the FOSS movement [10:26]. Recognizing mistakes that OSHW can avoid, and being cognizant of the differences between hardware and software can only mean good things for the movement. However open source hardware is only beginning.
We are approaching a singularity of sorts in our ability to create. Thanks to not only technological advances, but also to finding new ways of making existing technology available for lower cost (like OSHPark’s method of bundling PCBs together on a single panel), I can now create something that would have cost thousands or tens of thousands of dollars for a small fraction of that. The OSHW movement is completing the democratization of engineering that FOSS began.
Board photo courtesy of Laen Neal.
Richard Stallman image by Thesupermat (Own work) [CC BY-SA 3.0 (http://creativecommons.org/licenses/by-sa/3.0)], via Wikimedia Commons