Joshua Pearce on the Open Source Lab and democratising science
One of your books is titled “The open source lab”. What’s the big difference between an open lab to an non-open one?
The Open Source Lab is about using open source software and open source hardware in order to do scientific research. Normally scientists just purchase proprietary equipment or make it from scratch, but that’s very time consuming and challenging. On the other hand the Open Source Hardware mentality can be applied in the scientific realm. If one scientist develops a scientific tool he or she can share it with the rest of the community. And if its shared with all of its files, all of the electronics, the design files, and now with the advent of digital manufacturing, things like 3D printers, it becomes very easy for the next scientist to make a complete, perfect replica of this piece of equipment. So these hardware designs are shared with open licenses so the second person is obligated to re-share any improvements they make. And what we found is extremely rapid diffusion of innovation where we are creating much more complex, much more sophisticated, and all-around better scientific tools. They are able to be manufactured by the scientists themselves, very quickly and for far less costs then they would normally purchase proprietary equipment.
Your main argument is the enormous costs of scientific devices and how much money could be saved by using self made scientific tools. However, developing your own tools requires a lot of technical expertise, so do you think the educational system prepared researchers well enough so they already have this kind of expertise and if not how does it need to be reshaped to acquire said expertise and how much money would need to be invested?
Ok, now that’s a great question. With the digital design files, any scientist, even if they are not sort of you know- well versed in electronics — should be able to replicate the tool. You only need one scientist in the beginning to make the open source hardware device.
This entire concept is taking off and there are hundreds of designs now! However, there are thousands or millions of scientific devices. And so we still have plenty of work to do to get to all of them while driving the costs down by 90–99%. We do need improvements in education so that every scientist feels comfortable with let’s say soldering and using a 3D printer and using a laser cutter, so that they can more rapidly improve their own labs. Both in sophistication of the tools, but also radically reducing the costs.
In your paper about the Return On Investment in Open Source Hardware Development (available here), you not only outlining financial benefits but also how scientific research would be changed in other ways. How would increasing the return on investment of scientific research through the application of open hardware affect research in a broader spectrum?
So right now when I read one of the top scientific articles, often times the method section is the shortest and the most poorly written. And so it becomes difficult to replicate the experiment and sometimes that even is done on purpose in order to kind of keep other groups from competing on the forefront of science. But if we use the open source hardware method, as soon as a paper is published that hyperlinks back to that original piece of hardware, everybody in that subset of the scientific community can immediately catch up to the person who is in the lead. So this will accelerate all of science, because not only will they have access to know exactly how to do the experiment, they have access to be able to know how to make the machine and they will be able to make it, because it costs much less. And so rather than just have the high financed labs, the best labs in Europe and the US, being able to contribute to the forefront of experimental science, now everybody can. That means labs can participate in the developing world and that even means this works even at the lower education levels. So there is no reason high school students can’t be doing real science with top of the line tools with open source hardware being able to make it economically viable for them.
So it is interesting to take a scientific funder’s perspective — usually when they fund equipment for scientists they just buy the equipment for them essentially. But if instead they invest in the development of open hardware they get enormous returns on investment, because it basically costs the same if you develop a tool or if you buy a commercial one. But then the next generation, when everybody is just downloading and fabricating the tools themselves for let’s say 1% of the costs, that’s when the funder gets enormous returns on their investment. And so they will get a lot more science out for every dollar that they put in.
Gaston Bachelard emphasizes the strong interlink between scientific thinking and technology in modern science. If all scientists from tomorrow on were to solely use open hardware and self developed tools how would that affect their research?
So first of all everything will go much, much faster. So instead of sort of being limited by the relatively slow peer review process you can accelerate it so that many more people are being able to participate all over the world. If I read about the latest break through in Nature as I’m sitting around on a Saturday morning, I can immediately download the tool and work on it with my children — you know — in our home lab. And so I think this will be a lot more inclusive science, so that many more people can get involved in science. Science is actually fun and exciting when you get to do the real thing. Unfortunately, particularly in the educational system, that turns into cookie cutter projects, where you already know the answer — there is less fun in solving these problems like let’s say a jig saw puzzle where you already know that somebody knows the answer. The fun and excitement of science is getting to do something new and so if everybody can be at the ability to push out to the ledge, where the next experiment that you do will actually be something new and push all of science forward, that is very exciting! So we can hopefully pull more people into science that will be able to understand it better that will be able to move on it quicker and then the professional scientists will have enormous quantities of researches. Even if we maintain the same funding levels that we are at now we will be able to do so much more with it than we can currently. And so a lot of the problems that we have, and there is lots of them in the world, will hopefully be able to solved using science and technology, in a more rapid way.
Will there be a de-centralization and democratization in science?
Yes, absolutely. De-centralization and democratization, everybody gets to participate. Not just the super-funded.
Can lab equipment generated by DIY 3D printers ever meet the high standards and precision that large labels usually are expected to provide and if they don’t wouldn’t it harm research to use DIY equipment as experiments aren’t as precise or reproducible?
That’s a great question and I think that’s the majority of concern of the majority of scientists now. We need to do DIY scientific tools following the scientific paradigm correctly and have them calibrated and validated to be just as good as — if not better- than anything you could possible commercially buy. And in fact, in general open hardware tends to be better, because you can customize it to your own specific scientific needs for new experiments that no one has done before.
Thinking about where science funders need to go in the future — validation. To validate a tool is expensive. You have to run a bunch of tests on it, and you have to make a test that someone else can do in the field in the middle of nowhere, in order for them to be confident that they have a tool that is precise and accurate enough. And so all the tools that have come out of my lab, that’s what we do. That’s what the scientific paper that we are writing about is about. It’s about making the tool — and then making sure that its doing exactly what we think its doing. And if people follow our instructions to the T, they will get an accurate and precise tool that they will be able to match anything that is available commercially. And usually it will be able to do significantly better than any of what’s available commercially.
We do need funding for that. Just because something is open source doesn’t mean that its automatically necessarily better. You do have to validate and make sure that the scientific tools are doing what you think they are doing.
The one advantage of open source — because you are building the tool yourself and you have complete control of the source code and you know exactly what’s going into it, you have a very fundamental understanding of the way that it works and so you are never guessing. You never rely on a black box piece of software or hardware to solve your problem. You know exactly what its supposed to be doing. You know what its abilities really are. So if you take the example of our open source syringe pumps we know exactly how good they are — so if we are doing an experiment where you need a certain amount of fluid to put it into a system we can make sure that we get to that with the kind of resolution that we need. And if our system can’t do it then there is ways to improve upon it and all of these have been outlined in the open source hardware paradigm.
What are you currently working on with your research group? Do you work on a tool now?
Oh yes, always. So the latest fun things that we put out from our group is we’ve turned our 3D printers into scientific platforms. So in general we are using old-style RepRaps now. Usually they are off of a pulley system, but instead we put this directly off of a lead screw. And so this gives us a high resolution, they are not as fast as other printers go, but it gives us high resolution of movement and so we were able to do things as fluid handling, where you might have a hundred thousand dollar robot dispensing different chemical reagents in say test tubes, now we can do that ourselves using the 3D printer as a stage or platform for being able to do scientific tests. We also attached a microscope to it, and so you have a 3-D microscope stage. And again these are devices that normally cost more than a hundred thousand dollars and we were able to replicate it with the cost of a RepRap which is usually less than a thousand dollars, depending on how fancy we make it. And the cost of the microscope. And so we are taking this — the fundamental open source hardware components which are the micro-controlers and the mini computers and the RepRap 3D printer and turn those themselves into scientific devices.
Let’s see, we’ve always got a lot going on that is usually tailored specifically to the type of research that I’m doing. So our group is primarily a solar photovoltaic group — trying to turn sunlight into electricity. So right now we have an open source gimbal system that we are working on with a group in Canada. And this again is a 9000 dollar tool to be able to change the angle of something and what we are doing is we want to be able to change the angle so that we can look at the solar cell with different colored light with different angles to see if the new plasmonic structures that we are putting on top of it are actually behaving the way that theory predicts. If we are right the small metal nanoparticles will create plasmonic superabsorbers and improve the solar cell efficiency. And so we are making these ridiculously customized super tools for ourselves. When we are done then anyone else that might need the same type of system is of course free to use it and they only have to share the improvements that they make, which could also benefit our lab. We were one of the first groups to make a simple lab jack, and now there is a whole slew of different, much better lab jacks than our original one and that we use now any time we need one. And that’s a simple tool that everybody uses.
What we have been trying to do is sort of climb the ladder of sophistication to where we are getting now very sophisticated, very costumed tools for specific types of research. And we are not the only ones doing this, hundreds of scientists are also jumping on and there is literally libraries now of easily more hundreds of tools in every discipline in science.
Do you use thingiverse to publish these designs or which platforms do you use?
So that is a great question. So one thing we really like about thingiverse is the customizer app for OpenSCAD as our group uses OpenSCAD a lot. However, we also use Libre 3D and Youmagine. And for any of the scientific tools those should also be at the very least put on the NIH 3D Print Exhange. They have a separate section just for scientific tools and that’s where I think the more serious scientists head, but their database is still not quite as smooth as YouMagine, where its just easier to upload and do everything. So we usually cross post in a couple of different places to make sure we are hitting the widest possible scientific community. The other thing that we should be aware of is that Elsevier, which is the largest scientific publisher in the world, is coming out with a journal called “HardwareX” this month. And it will be an open access journal, where all the device specifications, bill of materials, software, hardware, all the designs have to be uploaded in order for the article to be published. And the articles are just about scientific hardware. And so we are hoping that that has a catalytic effect — that basically when you submit a normal journal article you also at the same time submit a “HardwareX” journal article to compliment it. So that you have a validated version of your particular hardware that happens to be open source to help other scientists replicate your work and build upon it. So it is literally becoming mainstream — you can’t get much more mainstream than that.