Since 2006 I’ve organized* Sketching in Hardware, an annual summit on physical computing tools. Forty of the smartest designers, developers, researchers, artists and educators I know get together for three days to talk about tools that make it easier to develop hardware products, responsive environments and digital art.
Hardware is still hard, but a lot has changed since I started the event:
- In 2006 we had at least five different, yet broadly similar, microcontroller breakout boards, including a new one called the Arduino. By 2009 the Arduino was what everyone talked about. By 2012 we were back to talking about five different boards, but each one occupied a different niche in terms of price, power and size.
- The early days emphasized one-off projects: prototypes, art installations, design explorations. By 2011 the conversation was as much about how to make 100 of an idea as it was about how to make the first one. By 2013 multiple participants, Sparkfun, Modular Robotics, Central Standard Timing, Sifteo, were engaged in serious manufacturing, with some (like Sparkfun and Modular Robotics) built their own factories, which brought in discussions of fabrication, funding and the sustainability of electronics.
- Education initially focused on getting university design and art students to explore hardware without help from trained engineers. In the last couple of years the conversation has evolved into include virtually every group of people without a formal technical education. Middle school kids, hobbyists, and experimenters with no formal education at all.
- The shape of the final product has also evolved. In 2006 a hardware sketch was a small box of wires, hidden under a table. By 2013 we had tiny wearables, smart clothing and systems that spanned whole buildings. This year we had an artificial aorta and a 3D printed edible raspberry. And of course boxes of wire.
Here are some trends and interesting new projects I saw in 2014, though it’s difficult to capture everything that happened at Sketching, especially late into the evening in Berlin’s biergartens:
- The Web is about to eat hardware. If it’s increasingly difficult to separate the Web from software, and software is eating hardware as per Andreessen, then soon the Web is going to eat hardware. Developing digital hardware looks more like Web development every year: boards that use node.js to create RESTful interfaces to small hardware devices that make them act like Web sites, such as Vlad Trifa’s presentation of EVRYTHNG or André Knoerig’s presentation on IoT Toolchains. Moreover, the development tools and approaches are beginning to resemble those used by Web developers, rather than embedded hardware developers. Phil Van Allen presented how he’s rebuilding NetLabToolkit from scratch using a modern Web approach. This is a major break from traditional embedded hardware methods, and carries with it a deep philosophical change: instead of writing proprietary firmware in C using closed toolchains and optimizing from the start, there’s a greater focus on building rapidly and cheaply with open source tools, storing everything in Github, and worrying about performance only for the areas that are clearly problematic. Two of the large chip vendors, Marvell and Intel, showed off their products that follow exactly this philosophy: Andy Carle showed of Marvell’s Kinoma, which embraces JavaScript, and Seth Hunter showed how Intel’s Galileo 2 board is based on node.js, Arduino and Web standards.
- We have answered “how” about hardware sufficiently well that we can start exploring “what” and “why.” It used to be that just making a thing that worked was a triumph, never mind what it did. Then making and shipping 100 things was a major success. Neither of those things is now easy, but the process has become sufficiently understood that people can ask bigger questions and use bespoke hardware to explore the effects of connected devices, not just their manufacture. Kate Hartman’s lab designs wearables that look nothing like a wristwatch while the New York Times’ Noah Feehan built a smart pin understands the topic of conversation.
- Our tools are getting more sophisticated. The Arduino family, the Beaglebone and the Raspberry Pi have addressed many of the problems people historically had in sketching in hardware, but they have also identified many situations where there aren’t yet good solutions. This year we saw several new tools that address situations where a generic microcontroller board isn’t enough. Norman Pohl showed hardware that hijacks a phone screen to add new ports, and an ultra low power board that runs for two months on a coin cell, while Travis Lee and Evan Shapiro demonstrated NOAM, an open source broker for synchronizing lots of different kinds of devices to create experiences that span scales and devices, similar to Spacebrew, which debuted at Sketching a couple of years ago. Both, incidentally, use a standard Web publish-subscribe model.
- Our tools have enabled us to think increasingly crazy thoughts. Of the applications that people showed this year, a couple may have not been possible to think just a couple of years ago. Erik Schweikardt is automating his robot-building factory with robots, and he’s open sourced his factory automation robot designs, because it’s not in his core business. Pedro Lopes decided that rather than just reading the state of our muscles for Quantfied Self purposes, he could put together a way to control our muscles, to puppet our bodies, with a smartphone and a piece of outboard hardware. What was the hardware built with? Open source sketching in hardware tools.
This list only scratches the surface of what we discussed and saw at the event, and I apologize to those attendees whose work and ideas I didn’t show (follow the link below to see everyone’s presentation slides). As always, though, I came away with more questions than answers. Digital hardware is becoming increasingly accessible, which is great, but its accessibility exposes assumptions about how embedded, connected information processing is made, who makes it and what it’s for. What used to be the realm of a few specialists working for giant companies with billion-dollar budgets is now accessible to many, but many of the infrastructures are still geared toward that old world. For this new world we need to think of new ways of funding hardware development, teaching it, distributing its products and understanding its cultural role. Desktop publishing in the 1980s effectively destroyed and reformed a printing industry that had existed for hundreds of years. Similarly sketching in hardware is likely to destroy and reform the 100 year-old electronics industry. What will this new world look like? At Sketching in Hardware 2015 we will hope to learn a little more.
Download all of the presentations here: https://app.box.com/s/hy7zh7450v7rsrcb4tci
More information about the event here: http://sketching-in-hardware.com/2014/
*The first couple of years I co-organized Sketching with Matt Cottam of Tellart and I’m highly influenced by the conversations we had in those early days. Thanks, Matt!
ADVERTISEMENT
