Being interdisciplinary
The February 2003 edition of MIT's Technology Review magazine contains a piece by Nicholas Negroponte entitled "Creating a Culture of Ideas". In it, he starts by describing the well-known environmental characteristics that spark innovation - heterogenous culture, taking the ideas of young people seriously, and tolerance of failure. Then he goes on to address education - where:
Expertise is bred by experts who work with their own kind. Departments and labs focus on fields and subfields, now and then adding or subtracting a domain. Graduate degrees, not to mention tenure, depend on tunnelling into truths and illuminating ideas in narrow areas.
[ "Creating a culture of ideas" - Nicholas Negroponte in Technology Review ]
So - although the lessons of innovation in places like Silicon Valley have been well documented, often by academics, the same lessons haven't been implemented in schools and colleges. Negroponte points out that in academia, the goals of "being interdisciplinary" and promoting "openness and idea sharing" are at the same time both "banal" and "almost impossible".
Later on in the same February 2003 Technology Review magazine, there is an in-depth article about supercomputering by Claire Tristram. The article starts with a description of NEC's Earth Simulator supercomputer (pictured here in David Bullock's vacation photos). The Earth Simulator uses a vector architecture (where processors request a list of instructions at once rather than retrieving them from memory one-by-one: see this slide by Gordon Bell) which dates back to the 1970s. The Earth Simulator outperforms supercomputers built using a newer architectural concept: that of clustering commodity machines together for parallel computing. So, unless we conclude that vector computing is the end of the road, more innovation is needed.
Next, the article describes two innovative projects which are rethinking computer architecture. One is the Gilgamesh project at Caltech (described here) for the arrangement of memory and logic together on a chip, reducing travel time for data and instructions. Then there is Cascade:
Another option is simply to turn the entire architecture on its head. Cray's Burton Smith and Caltech's [Thomas] Sterling are cooperating on a DARPA-funded project they call Cascade. The two are investigating ways to exploit the fact that in high-end scientific computing, the heft of the data alone is often far greater than the heft of the applicaiton program. In other words, if the stuff stored in memory is so much larger than the program needed to run it, why move it at all? Why not move the program to the memory instead?
[ "Supercomputing resurrected" - Claire Tristram in Technology Review ]
So - how is this computer architecture research related to "being interdisciplinary"? The answer is that cognitive psychology and neuroscience clearly have lessons for high-performance computer architecture - in the brain, processing and data are not seperated. Processing occurs where the "data" is. I was lucky enough to study mathematics and psychology at Trinity College Dublin, and I remember days when I would attend seperate lectures on microprocessor systems, cognitive neuroscience, and the cognitive science of thinking. It was fascinating to compare computer processing with brain processing, and computer memory with human memory - I remember the thrill of discovering the books of Douglas Hofstadter, which link the two fields together. Hofstadter's ideas also support interdisciplinary studies, since "pattern perception, extrapolation, and generalization are the true crux of creativity" [Fluid Concepts and Creative Analogies].
Studying for a double major is a simple, practical way of "being interdisciplinary" which is increasingly popular, according to the New York Times. It means that even when academic departments do not communicate with each other, as Nicholas Negroponte accurately describes, you can still benefit from a combination of ideas. That year at Trinity, only two of us chose the mathematics and psychology combination [I guess I sold out after college, but John is still going strong in neuroscience research]. I'd recommend a double major to anyone thinking of starting, or going back to, college.
|
