Reinventing the Scientific Method
A Techwise Conversation with the author of a new book, Reinventing Discovery
Steven Cherry: Hi, this is Steven Cherry for IEEE Spectrum’s “Techwise Conversations.
The Internet and social networking are transforming entertainment, journalism, retail commerce, and politics. And more slowly but still surely, they’re changing the way science and engineering research are being done.
My guest today knows scientific research from the inside and decided to write a book about the ways science has and hasn’t changed. Part descriptive and part prescriptive, his book, Reinventing Discovery, was published by Princeton University Press last month. It’s a provocative look at the modern process of invention.
Michael Nielsen is a theoretical physicist specializing in quantum computation and quantum information—in fact, he coauthored the definitive textbook on the subject. He most recently was a senior faculty member at the Perimeter Institute for Theoretical Physics, a Canadian research center started by one of the cofounders of Research In Motion, the makers of BlackBerry smartphones. Michael, welcome to the podcast.
Michael Nielsen: Thanks for having me, Steven.
Steven Cherry: One thing you write about is the way that the Internet and social networks have made it possible for professional scientists and amateur scientists to collaborate. Tell us about the Polymath project.
Michael Nielsen: The Polymath project was this experiment done—I guess it was started by a British mathematician named Tim Gowers, who’s one of the world’s top mathematicians, and he’s also a blogger. And he decided to use his blog to attack a problem completely in the open and invited other people to contribute if they wanted to. And the experiment actually turned out well—at least the first version turned out to be very successful, with a large number of contributions coming from many different people—and they quickly cracked this open problem.
Steven Cherry: Very good. You know, you’ve written that some remarkable amateur science is being done nowadays because amateurs have access to the same data that scientists do. I guess the Sloan Digital Sky survey is based on that idea?
Michael Nielsen: Yeah. So what the Sloan Sky survey did was they built this robotic telescope in New Mexico—actually a little town called Sunspot—and it’s been slowly scanning the sky taking photographs. And of course, they have a very, very large number, and it turns out that some of the time it’s best to have human beings analyze those photographs directly—not to have it done by computers. And one of the most effective ways of doing that has been through a project called Galaxy Zoo, which has about a quarter of a million amateurs to look through the photographs. And so far they’ve done about 150 million galaxy classifications based on that.
Steven Cherry: You know, it occurs to me, just to pick some examples, Joseph Priestly, the man who discovered oxygen, was in fact a priest; Blaise Pascal, who basically invented probability theory, never made any money at math or science except for his winnings at the card tables; Ben Franklin was a diplomat and a postmaster and a publisher but also the first electrical engineer. Isn’t this in some ways a revival of the old method of science, when amateurs were among the leading scientists of the day?
Michael Nielsen: In some ways it’s definitely true that in large parts of science, there’s a real advantage to having a lot of expertise. It’s pretty hard to make a contribution to quantum field theory without a fair background. With that said, some of these tools sometimes—particularly when they can aggregate contributions from large numbers of amateurs—really do let them do things that the professionals just can’t on their own.
Steven Cherry: Science was working pretty well for a couple of hundred years, at least. I mean, does it really need to be reinvented?
Michael Nielsen: Well, of course, I have no problem with science; it’s something that’s obviously one of the great achievements of our civilization. With that being said, obviously new ways of doing things are wonderful and much to be desired. I’ll give you an example: One doesn’t need to think that personal computers or laptops have anything wrong with them to be very keen on the latest iPhone or iPad or whatever—it’s a new approach.
Steven Cherry: One thing that seems to be changing is peer review.
Michael Nielsen: Yeah, I think at present if you’re talking about publications in peer-reviewed journals, then this is something that’s been fairly stable for a few decades. It’s actually a relatively modern invention, more modern than people think, in the sense that it only became the standard in science by the middle of the 20th century. I don’t think that’s necessarily going to change all that much with new online methods of discovery. I think people will for many decades to come still use peer-reviewed journals, but it will be complemented by other ways of doing things. To come back to this example of the Polymath project that we talked about early on in the interview, one of the interesting things about it was that this whole process of collaboration was done out in the open, and so in some sense there were actually a lot of opportunities for mathematicians to check over what other people in the collaboration were doing. So it was a very transparent, very easily checked kind of a process and so in that sense almost a slightly different style of doing peer review. Now with that said at the end of the day, they still published some peer-reviewed papers and said that was still the final outcome, but early on there was only a slightly different process of checking than is ordinarily the case.
Steven Cherry: Michael, you know, most scientific discovery in the 20th century was located in universities, research institutes, corporations. I take it you think discovery is going to be taking place in different places this century.
Michael Nielsen: Yeah, certainly it’ll become much more distributed, much more spread out. There’s an interesting question, which is, to what extent will it start to be seen outside the traditional academic places? You think about this example of Galaxy Zoo, which I talked about before, you know it’s a quarter of a million people. Admittedly, a lot of those people are only peripherally involved, but even if you’ve got a few thousand or tens of thousands of really hardcore users, that’s a lot. You know it’s probably more than the total number of professional astronomers active in the world, and they’re from all over the place, so that’s a really interesting phenomenon, and nobody knows what it will lead to.
Steven Cherry: You know, I mentioned at the top that your book is part descriptive and part prescriptive. What are some further changes that you’d like to see wrought by the Internet and social networking?
Michael Nielsen: Well, there’s a lot of very promising ideas that seem like they should work in principle and yet they don’t work in practice. One example that I give in the book, for example, is a lot of people have tried building scientific wikis to collaboratively build knowledge bases about the latest research accomplishments. And often these wikis haven’t done quite as well as you think they should, and part of the reason, of course, is that there’s a real opportunity cost involved in contributing to a wiki: Should you—particularly as a young researcher—should you spend your time doing that, or should you spend your time writing, working towards peer-reviewed scientific papers? And from a scientific career point of view, the answer is of course pretty simple: You should work on the scientific papers, because there’s not going to be much credit for you if you adopt these more radical, newfangled tools.
Steven Cherry: Are there any other examples?
Michael Nielsen: Two very big examples, very broad examples are data sharing and code sharing, both of which are things which in most disciplines people don’t get a whole lot of credit for. Very often there’s a lot of very important scientific knowledge locked up in, for example, code, which you might use to do all sorts of data processing in the laboratory or simulations or whatever. And yet, when I talk to people who write a lot of code as part of their scientific job, they’ll say that very often they’re extremely reluctant to release that code publicly because well, first of all, it’ll end up being a pain; they’ll have all sorts of support and maintenance requests from other people. And second, it’s not something they can use as part of their tenure case; it’s just not something they get a whole lot of credit for.
Steven Cherry: So really it seems like academia has to catch up with the Internet and start rewarding behavior that’s pretty useful to society.
Michael Nielsen: I certainly think so. A phrase that’s really stuck in my head is—a biologist commented to me once that he’d been sitting on a genome for an entire species of life for more than a year. And he’d been doing this because his collaborators didn’t want him to share that data, didn’t want him to upload that data online where other people could use it. This, of course, is really not very uncommon within science, and it seems like really a tragedy, a lost opportunity. Other people could potentially have made all sorts of useful discoveries with that data.
Steven Cherry: Very good. So in a “physician, heal thyself” way, how is your own work been impacted by the changes you write about?
Michael Nielsen: So the largest single open project that I’ve done, not surprisingly, is writing my book. It was a four-year-long project, and a funny little thing—which was the most useful thing, ultimately—was actually just using a public bookmarking service to share the research that I was doing. So when I’d see something that was interesting, an interesting paper or interesting news article or blog post, I’d bookmark it so it would be publicly shared, and then there’s a discussion site called FriendFeed, which would allow little discussions to take place around that. And at first not very many people were commenting, but ultimately it’s this little community of people who see a lot of these bookmarks and provided in some cases really remarkable extra information. They’d debunk it or they’d follow it up or say, “You should be looking at this or that or whatever.” And I think I ended up sharing about 1500 or so links there, and I got thousands of comments from hundreds of people, and certainly the book’s a lot better than it otherwise would have been as a result. And I used the same pattern, in fact in all of the research that I do now.
Steven Cherry: Very good. Well, that sounds really worthwhile. Michael thank you so much for talking with us today.
Michael Nielsen: Thank you so much, Steven.
Steven Cherry: We’ve been speaking with Michael Nielsen, author of the new Princeton University Press book Reinventing Discovery, about how the Internet and social networking are changing the way scientists do science. For IEEE Spectrum’s “Techwise Conversations,” I’m Steven Cherry.
Announcer: “Techwise Conversations” is sponsored by National Instruments.
This interview was recorded 17 November 2011.
Follow us on Twitter @spectrumpodcast
Segment producer: Barbara Finkelstein; audio engineer: Francesco Ferorelli
NOTE: Transcripts are created for the convenience of our readers and listeners and may not perfectly match their associated interviews and narratives. The authoritative record of IEEE Spectrum’s audio programming is the audio version.