Ben Fry is the director of Seed Visualization in Cambridge, Massachusetts, a company that uses coding to help visualise and understand complex data. As a doctoral graduate of John Maeda’s Aesthetics + Computation Group at MIT he is also, with Casey Reas, the co-creator of Processing, the open source computer programming language that has enabled artists and designers to sketch and make work with code. Newly installed at Seed and its experimental Phyllotaxis Lab, Fry talked to us about his work, which has ranged from illustrating the human genetic code, mapping the 26 million roads in America, and even contributing to the look of some of the most innovative sequences in the Steven Spielberg film, Minority Report. At 34, Fry is already the recipient of numerous awards for his interactive design work and Processing 1.0 – not yet even a year old – is used by tens of thousands of people every week.
CREATIVE REVIEW: You started working at Seed Visualization in March this year. Can you explain what Seed does, alongside its Phyllotaxis Lab? It seems that in terms of what you’ve been working on, from being at MIT onwards, that you’re now at the perfect place for your skills and interests?
BEN FRY: Seed Visualization is the client side of things, the Phyllotaxis Lab the research and internal side. For example, for the World Economic Forum meeting in Davos in January, we developed a piece that depicted connections between the different sessions at the conference – there were 180 of them – through various keywords found in their descriptions.
General Electric was another recent client where we wanted to show how the company was involved in healthcare and to show the kinds of things we can learn from healthcare. [The online project visualises the links between gender and age, and a range of illnesses such as heart disease and diabetes].
For the Lab side, this is where we want to push the state of the art a bit, and pursue our interests without the constraints of a client project. The idea is that research and speculative work helps feed ideas for other projects. In the past I’ve found that if I work too much on the practical side, or too far on the purely aesthetic side, the results really suffer. You have to pursue a mixture of both. My Disarticulate piece – a pair of prints, one of which showed the results of a virtual interpretation of another art piece by Casey Reas – came along during one of these periods. It also sometimes coincides with when I’ll work on a show or exhibition of some sort.
CR: What’s the overall aim, or purpose of visualising all this data?
BF: The personal answer is that it’s just what I’m curious about. I enjoy finding things out, building things that help me make sense out of a complex idea, and being able to see that once it’s done. So for instance, if I read about the H1N1 ‘swine flu’ virus, how it has travelled to 89 countries, and how the virus has continued to evolve, I immediately want to start building something that actually shows that phenomenon. How much has the virus evolved? What does that look like on the dna level? Can we track its changes across the globe? How can we see the progression of the pandemic? I’ve been lucky to be able to pursue this sort of thing and make a living from it.
Professionally, companies and governments are ever-more inundated with data, and need help making sense of it. It’s been fascinating watching the progress over the last ten years since I started lecturing more frequently: I no longer need to make the case about ‘lots of data’ or how visualisation might help. The clients are already asking for visualisation help, because they’re completely overwhelmed.
CR: You’ve described the process of visualising data as a kind of habit? Is it one you’ve had for a long time? How does the challenge of visualising data manifest itself now?
BF: Yes, at least for the last several years, as I figured out that this is what I wanted to do, and found or developed the tools to support it. I have a tendency whenever I hear about numbers or a good story about them to start thinking about what they look like, and if it makes me curious enough, I’ll try to build something that tests the idea.
For instance, a friend who works in genetics was telling me that she thought Darwin was full of it, and that he had stolen most of his unique ideas from another scientist named Wallace. Wallace wasn’t as high society as Darwin, and was just happy to have some of his ideas presented by someone of his stature. She cited as proof that the first edition of his Origin of Species was a kind of incoherent mess, and it wasn’t until later editions that he really figured out what was going on. Phrases like ‘survival of the fittest’ didn’t appear until the fifth edition. This got me thinking about how Darwin’s theory of evolution – usually thought of as a monolithic and staid idea – had in fact ‘evolved’ over time.
More recently I began building software that shows the evolution of the text from one edition to the next, as it grew from 160,000 words to 190,000. Interestingly, the process has had the opposite effect of what I expected: instead of thinking that the guy was a flake, I’ve come to have far greater respect for the book and his ideas.
CR: Clearly there is a mind-boggling amount of data out in the world today, particularly on the web. Are there schools of thought as to how best it can all be managed, collated and displayed? What’s going to happen next?
BF: I think it’s just about more, more, more. This is my job security: people talk about how we’re inundated with ever more data on a daily basis. This gets tiresome, since it doesn’t really mean anything – it’s like having more channels on television. But the bottom line is, this is only going one direction: there’s no trend toward ‘less’ data. That’s a fascinating thing, if you think about it.
On the downside, the transience of data and ubiquity of cheap storage makes hoarders out of all of us. Look at what it’s done to Google, or any number of companies who have had problems with security breaches and customer files and personal information being accessed.
In half the cases, why did they even still have that information around, years later? Because it’s easy to store and collect, it seems like a good idea.
CR: Going back to your time at MIT, what you learned and experienced there has no doubt impacted on where you find yourself at Seed. What did MIT teach you?
BF: Our advisor, John Maeda, pushed very hard for our careers – he was very different from many other advisors in that regard – he was obsessed with us getting ahead rather than using us to get his work done or push his own agenda or career. That was quite a gift – it opened a lot of doors for us.
One of my favourite things about MIT is that it’s a culture of people who like to make things. To a person, people have one or several things that they’re freakishly good at or smart about. I get a charge out of being around people who are smarter and more talented, so this was a lot of fun.
Career-wise, I think the biggest thing was that it was a place where I could find my voice, whatever that means. In an environment of exceptional people who do esoteric things, you’re free to pursue your own odd path and it seems quite mundane and normal.
CR: And post-MIT, you’ve worked on the kinds of ‘tools’ that would enable you to visualise human and animal genetic data – what kinds of tools were these?
BF: This was a handful of things. I did a lot of work with haplotype data [the forms of genes transmitted in chromosomes] during my Ph.D., which is all about tracking variation in genomes and understanding how one person’s genome compares to the next. One of the main themes was this idea of what we can do now that we’ve collected many genomes – whether of humans, or being able to compare humans with lots of other animals.
Completing the human genome isn’t nearly as interesting – or useful – as completing the chimp, mouse, dog, fruitfly, and yeast genomes as well, and being able to compare them to understand what makes us all tick.
CR: In terms of the Processing language you created with Casey Reas – as it’s something that is continually updated and added to by its users, has it evolved as you imagined?
BF: Its evolution in terms of features comes primarily from people who contribute ‘libraries’ that extend what it can do. The core of the software gets less input, it’s all about how people create simple extensions that support some activity that they had wanted to do but couldn’t. For instance, our audio support was terrible for a long time, and a handful of people developed libraries to improve that. More recently, Damien de Fede developed a really solid audio library called Minim that we now include with the software for people who want to use it.
I think the software has gone through significant evolutionary steps, the alpha, beta, and 1.0 releases probably should have been 1.0, 2.0, and 3.0, but Casey and I are sort of perfectionists so we weren’t ready to do that. In these, the community has sort of pushed – shoved, really – us along to the next step as they wanted to do bigger and broader things with the software.
For the future, we’re mostly trying to improve what’s already there. On one hand that means making things like graphics and video work better – run much faster – and on the other it’s about making the environment easier to use: better editor, better error messages. Not terribly sexy but it will have a significant impact for our core audience.
I don’t think we had a clear idea of how things would evolve, really. I don’t know if we expected it to turn into a professional development tool – graduating from its roots as a teaching and sketching tool. I don’t think I expected that I’d be developing nearly all of my work with it a few
CR: Do you see Processing as essentially a ‘simple’ system for designers to use? What were yours and Casey’s original intentions with it?
BF: A large part of it was making things simple that should be, and making the hard things tolerable or at least accessible – is that Einstein’s notion? We want things to be simple when we’re sketching because there’s no need to rewrite the same code over and over because it just gives us rsi. For beginners, they don’t need to type those lines of code either, because they get in the way of what they’re trying to do.
CR: So in this sense, was it consciously looking to get closer to how people work with a sketchbook and pencil? A way of getting nearer to the pure creative idea?
BF: Absolutely. We’re still a long way off because we’re talking about a fairly traditional, text-based programming environment. But the concept of trying out ideas, to be thrown away, or evolve into something else, is the absolute core of what we’re aiming for. Sketching is so important for developing ideas, and we wanted to use that approach when working with code.
CR: In terms of your working day, you’ve said that you’d like computing to feel more like working in the studio, like walking, or drawing in your sketchbook and that ‘gesture-driven’ interfaces, like the ones featured in the film Minority Report, might help this to come about. How far away is this kind of work as a reality?
BF: Well the guys at the LA-based studio, Oblong, have their system up and running, and it’s a bit shocking to see it actually work.
Their co-founder, John Underkoffler, was part of the Tangible Media Group at MIT, which was adjacent to ours on the fourth floor. We spent a lot of time working together and, after finishing his Ph.D., he worked as Science and Technology Advisor for Minority Report. The gesture interface in the film builds on his Ph.D. work. He also grabbed a couple of our projects to use in the film as part of other sequences.
After the movie, John and some others went to build the system for real, since they knew it could be done. So Oblong is their start-up that’s making it all happen. I spent a couple months working for them on projects a year or two ago, though sadly I don’t have any good documentation of the work to show off.
Of course, with ‘gesture-driven’ interfaces you have to build all sorts of applications to support this new method of working – think of the lack of applications shipping with the original 128k Mac – before you can really put the system into use more broadly.
One of the missed points about using ‘gesture’ is that you’re not going to use all your old applications and retrofit them for a new interface, you have to rethink things entirely. You won’t see Microsoft Word running in such a system, it doesn’t make any sense to do text editing while flapping your arms.