Is Raspberry Pi homebrew the right way to do science?

[Robert Blackwell]

At a recent academic event, we were shown a UAV (unmanned aerial vehicle, drone) that had been put together by a student to experiment with a particular kind of multispectral camera. I saw it as a quick and dirty integration using off-the-shelf parts, very much in the experimental scientific tradition of Faraday and others. It was cheap, functional and another good example of Raspberry Pi for prototyping and proof-of-concept work.

Surprisingly, we were told that this kind of work is better done by specialist engineers and that the student had “wasted his time” when he could have been doing science. It seems that he spent nearly a year on it. The way forward, we were told, is for scientists to propose ideas and encourage engineers to work on them. Not since the days of Gauss could a person know everything; gone are the days when a single person could have all the necessary skills.

I think this is not only wrong thinking, but damaging and it led me to the following questions:

1. Why would science not want to exploit the trend in widespread, cheap, modular, easy to use, kit-style, open electronics?
2. Why did it take a year and not a fortnight to build?
3. Is this teaching widespread and is it one of the reasons why it's difficult for UK start-ups and early stage businesses to hire hackers?

If you're a programmer and you've ever built anything of your own on Raspberry Pi, you don't need me to tell you how much fun it is. A few years ago, building hardware was complex and had long lead times. Now, a quick Python script and some plug in modules allow rapid integration, proofs of concept and minimum viable products that are robust enough to stand up to real workloads.

Go along to a Raspberry Pi event or look on YouTube and you'll find geeky 12 year-olds building amazing things over the weekend. Some of it is gaffer-taped, but much of it includes 3D printed and home-manufactured parts. Presumably the above project took a long time because the student didn't have ready access to the skills and support he needed.

The UK economy is facing challenging times. We don't know what effect Brexit will have and whether large companies will stay and grow. The only sure way to job creation is starting new businesses and that requires entrepreneurs and hackers. (If you think a hacker is somebody who breaks into computer systems then please read Paul Graham and others and take a look at the YCombinator success and philosophy).

Hacker skills include:

• Ability to learn new things fast, not just through books but by hands-on experimentation, making stuff and taking things apart.
• Excellent facility with data and manipulation through software.
• An ability to construct software and hardware systems rapidly that are just good enough for the task at hand.

I only aspire to these skills, but I have been fortunate to work with, and hire, some talented hackers over the years. Unfortunately, they are few and far between.

UK education thinks a lot about STEM (Science Technology Engineering and Mathematics). There are many challenges here, but one of the greatest seems to be at the Science and Engineering interface. Rather than segmenting scientists and engineers, perhaps we should consider what it would take to engender more hacker skills?

(My thanks to NEXUSS students for reviewing earlier drafts.)