Pasteurized computing: the relationship between academia and industry

James Pallister

I recently attended HiPEAC’13, an international conference covering many aspects of computer science. Unlike some academic conferences, industry was well represented and catered for. On the first day of the conference there was a student poster session (which I took part in), followed by an industrial poster session the next day. There were keynotes given by both academics, and researchers from Microsoft Research (MSR) and KALRAY.

One of the important things that meetings between academia and industry allow is the transfer of knowledge both from universities to businesses and (less considered) from businesses to universities. This allows businesses to make use of technology on the cutting edge, while giving academics feedback about how their research is actually used, as well as identifying possible future markets for their ideas.

Most of industry is about applying research to get results, most visibly in product design, but also in many other areas. This kind of applied research is often different from what is done at universities – universities often deal with knowledge for its own sake. Both have their own place in the world, but when both combine the most interesting things occur. This phenomenon is well described by ‘Pasteur’s Quadrant’.

Pasteur's Quadrant

The diagram (shown to the right) categorises the type of research performed, based on the orthogonal aspects of fundamental understanding and consideration of use. Research for knowledge’s sake is contrasted with invention – purely towards creating something new. The Wikipedia page¬†describes it well, suggesting that in engineering research is augmenting utility and knowledge simultaneously. Louis Pasteur himself was the microbiologist who invented pasteurisation, as well as furthering the understanding of microbiology to prove the germ theory of disease.

I believe it’s meetings like the HiPEAC conference that help this: bringing both industry and academia together to share results of collaborative research. In particular we usually think of information flowing from academic research to businesses. However the opposite point was adequately summed up by a question asked at the end of Doug Burger’s (MSR) keynote: ‘What do you want us (universities) to research?’. His answer to this was a bit vague – no company wants to give away their earliest research plans for fear of being pre-empted – but a few days later I came across a blog post. This post, from Matt Welsh at Google, talks about ideas that Google has an interest in getting the academic community to explore.

Embecosm and the University of Bristol recently collaborated on a piece of research into compiler optimizations and their effects on energy consumption, which was presented at the PEGPUM¬†workshop and which we’ll be talking more about in future blog posts. Hopefully through more collaboration, the type of research we do can fall into Pasteur’s Quadrant, being both useful and furthering our understanding.

James Pallister is a PhD student in the Department of Computer Science at Bristol University. He is also a technical adviser to Embecosm and contributes to our research into the impact of compilers on energy consumption in embedded systems.

Leave a Reply