IEEE Spectrum Magazine, Sept. 2003
Is industrial research in electrical engineering withering away? I have bad vibes about this. As part of an ongoing study in which I am participating, I did a survey of publications in the IEEE Transactions on Communications over the last 30 years. Were the authors writing the papers from industry or academia?
In 1970 about 70% of the papers in these transactions were written by authors from US industry. For graduating students with aspirations to do research in those years the place to be was in one of the great industrial laboratories – Bell Labs, IBM, RCA, GE, and others. The research journals were full of papers from these famous labs.
Then something happened. A steady decline in industrial research papers began. Those great names started to disappear from the transactions. By 1990 only 15% of the papers in the IEEE Transactions on Communications were being authored by US industry. In order to get a more recent data point, I examined the papers from this year’s IEEE International Communications Conference (ICC ’03). In that conference only 7% of the papers were authored by US industry. Slightly fewer papers came from non-US industry, while all the rest of papers came from academia. Within the academic papers, a slight majority came from universities outside the United States.
Thus the decline in industry papers has been offset by an increased number of academic publications. There is no evidence to suggest that fewer research papers in total are being written. I have no data on numbers of submissions, but the journals and conferences are still filled to capacity, and the rejection rates seem to be rather constant. It’s just that now the research papers in communications come from the universities rather than industry. Those papers are mostly being written by graduate students who will have a hard time finding jobs in the US in these difficult times, but that’s another story for another day.
I’m sorry to see what is happening to industrial research. Labs are disappearing, and for those that remain, the sizes of the research organizations are being reduced. But that is only the visible part of the story. Less evident , but perhaps more important, is the fact that time scales for research have been shortened to focus on the near term, and that research projects have been forced to closely align with development organizations. Modern project management is being applied to research to decide which projects get funded and what milestones must be passed. To my taste, research is being turned into development. Business leaders see this increased control and management of research as beneficial. Perhaps they’re right. I don’t like it, but I don’t have to run a company.
The problem is, of course, that companies have to make a profit, and do it at a pace that satisfies the investment community. Because of its inherent unpredictability, “unfettered” research cannot be counted upon to contribute to corporate profitability. Research managers point to their calculations of past returns on investment in their research, but the business managers disbelieve the calculations. “What have you done for me lately?” is the unspoken question that hangs over any discussion of the historical benefits of research. It’s a tough world out there.
So it seems that research increasingly will be the province of the universities. But the journals and conferences are still full of papers, so what’s the problem? My theory is that without a grounding of reality supplied by industry, academic papers can tend to drift off into imaginary spaces. I think I see this happening now, as many papers seem to be minor variations built on top of other minor variations. With the pressure to publish in universities being what it is, such behavior is quite understandable.
During the dot-com boom, when venture capital was easily obtained, a number of leading universities were hotbeds of entrepreneurial activity. Most of the professors, and many of the graduate students, were especially well motivated to find practical problems. In those circumstances, I believe that the university research system functioned well, and in some ways perhaps better than industrial research. But those days are gone, and few universities have the motivation or the ability to find problems of practical significance.
Industrial research is not a perfect form either. It can become too grounded in reality to contemplate those elusive “out-of-the-box” solutions. Perhaps a combination of academic and industrial research best serves the advance of our technology. But one of those partners is being lost, and I see little hope of its resuscitation.
I had the wonderful experience of working at Bell Labs during those fertile, unfettered years. I’m sorry now that others are deprived of this opportunity.