An article in IHT today precipitated this post although the general idea has been fermenting in my mind for several weeks. This IHT article, “Young economists search for relevance”, describes how economics as a field is seeing a creative renaissance. The economics of today is more soulful than dismal, and more practical than ever before. It is diving into new areas like causes of racial inequality, high rates of HIV infection, benefits of television for some toddlers, how to maximize happiness, etc. As if to illustrate this trend, The Economist carried an article a couple of weeks back about “Economics discovers its feelings”.
All this branching-out is taking economists into the domain of epidemiologists, public health workers, psychologists, etc. and is expanding their influence beyond the traditional spheres of monetary policy and the like.
What has all this got to do with the computer industry? Well, if you buy that the network is the computer then you’ve got to be in awe of the dramatic changes that have taken place in the last 20 years. Who could have imagined in the early days of ARPANET that Internet would be about social networking, online gaming, blogging, ecommerce and entertainment? What a distance we have covered from the days of ftp.
Much of this change has not been predicted. After all, Bill Gates and Microsoft woke up to the Internet potential only in 1995. This is not surprising. As they say predicting innovation is something of a self-canceling exercise: the most probable innovations are probably the least innovative. P.J. Rourke writes in an article on innovation (subscription required) that the…
…history of humankind’s development can be summed up as the story of surprise. Adam Smith failed to forecast the Industrial Revolution despite his friendship with James Watt, inventor of the steam engine that powered it. And who would have prophesied MySpace, Oprah, or a TSA ban on hair-styling gel in quantities greater than three ounces?
You are probably still wondering where I am taking you with all this. My point is that the kind of soulful revolution that economics is seeing today is also strongly needed in computer science. I am not suggesting that the “old” computer science of operating systems, programming languages, database systems and networking is passé. No, instead, my view is that we need to build on the foundational subjects to address the new research questions that have suddenly become important. For instance, we need computer scientists thinking about moving beyond passive retrieval of documents to proactive supply of information based on a clever extraction of intent. We need computer scientists bridging community science and computer systems. We need computer scientists applying AI to the web. You get the idea.
In essence, this is about having our best computer scientists collaborate with experts in other fields, mostly in the realm of social sciences, to develop new insights and knowledge. This is a fairly radical change to the traditional computer science world of mathematics and algorithms.
Happily some of this change is starting to happen. Yahoo has recently added Ron Brachman from Darpa, Raghu Ramakrishna from University of Wisconsin (Madison), Michael Schwarz from Stanford, Andrei Broder from IBM Research, etc. to address these new inter-disciplinary problems of today. In Aug’06, Wall Street Journal covered this story in its article, “Hoping to Overtake Its Rivals, Yahoo Stocks Up on Academics” that resulted in some discussion here and here. I did a quick search and found that ACM is also rethinking the computer science curriculum. All this is good news indeed. We do seem to be moving towards a new field of computer science with a soul!
[Some days back I wrote about a similar change that’s needed among the software practitioners in my post, “A Whole New Software Engineer”.]
CS has as you know always had an applications aspect and a pure aspect to it. Used to be that navel gazing was at least equally, if not more, respected that reaching out. Over the last 10 years or so, academic research has migrated to emphasizing the multidisciplinary side. Our language to describe this emphasis has evolved: used to be inter-disciplinary, then became cross-disciplinary, and more recently we say trans-disciplinary: The idea is simply that it’s not a union or intersection of your ideas and mine, but that working on the boundary yields new ideas for both you and me. This trend is of course broader than in CS alone, but in the CS context, we clearly are living much more with applications.
(Incidentally, the incentive structures and the organizational structures are indeed changing to facilitate this change, for instance the “department” is less fundamental as a unit than it used to be. But measuring “trans” work is hard because there are few expert judges and few outlets for discussion at this point. And, of course, fundamentals are no less important today than they ever were, except I sense that thinking about fundamentals risks becoming somewhat narrowed by the focus on particular domains. What I say is contradictory at some level, fundamentals must be fundamentals, but my point is simply that it is becoming harder to be really pure.)
Another, possibly correlated, trend has simply to do with scaling. One can ask finer grain questions simply because we just have easier access to more data and control information. So, the experiential/intuitive guru of the 70s is being replaced by the “I discovered/validated this in the numbers” economics/business/domain expert. It’s hard to argue against numbers (unless you have another set of numbers from somewhere else
. But it does look like business, like economics, is increasingly influenced by this fine grain data analysis. My own interest is wireless sensor networks is along these lines: to see/control phenomena at a greater resolution than what was previously possible. Perhaps this trend makes us more efficient, but more interestingly it lets us pose and solve problems that we could not consider in the past.