September 5, 2013: People of ACM: Alfred Z. Spector
Thursday, September 5, 2013
Alfred Z. Spector has been Vice President of Research and Special Initiatives at Google since November 2007. Previously, he was Vice President of Strategy and Technology within IBM's Software Group, and prior to that, he was Vice President of Services and Software Research across IBM. He was also founder and CEO of Transarc Corporation, a pioneer in distributed transaction processing and wide area file systems, and was an Associate Professor of Computer Science at Carnegie Mellon University, specializing in highly reliable, highly scalable distributed computing. While there, he served as doctoral advisor to Randy Pausch and Joshua Bloch. He also served on the Markle Task Force on National Security in the Information Age that focused on how to use information and information technology to improve national security while protecting traditional civil liberties.
Spector received his Ph.D. in Computer Science from Stanford University and his A.B. in Applied Mathematics from Harvard University. He is a member of the National Academy of Engineering, a Fellow of ACM and IEEE, and the recipient of the 2001 IEEE Computer Society's Tsutomu Kanai Award for work in scalable architectures and distributed systems.
How have your experiences as a prodigious researcher as well as a top executive and successful entrepreneur informed your vision of computing in the information age?
Computing is full of opportunity in the broadest sense. I've done a bit of almost everything in my career (from microcoding to sales management!), so I feel our field is full of wonderfully diverse challenges. And, most of the challenges are even tractable. No dismal science, this! I find it strange that some still think of the field as one-dimensional, but that view is happily changing: witness the super-good enrollments in the newer-style introductory classes (e.g., Harvard's CS50).
Additionally, as our field is central to so many others (as I say, "∀, CS+X"), it's arguably even broader. At Google, with our mission to "organize the world's information and make it universally accessible and useful," we depend on innovation across almost all computer science disciplines, so I see the diversity daily, and am thankful I've witnessed a lot before assuming this job.
In an era of data-driven research and instant dissemination of information to vast numbers of people, what do you foresee as the future of scientific, scholarly publishing?
I emphasize to researchers and engineers that we should judge ourselves by our true impact. In a field that has a very significant engineering component and enormous societal value, we computer scientists are blessed with many routes to successful impact: we can release products that set a new state of the art. I believe that a great product from us or other institutions is very often a great contribution to the research state of the art. Alternatively, we can create new standards, open source code distributions, and data set releases. Consider the value of publications like the HTML 5 Specification or Android's source code releases. These are broad shoulders indeed from which many have seen a far more distant horizon.
But please note, I always emphasize that communication via scholarly, peer-reviewed articles (whether in conferences or journals) is also a very impactful and important activity; indeed, Google research scientists and engineers publish hundreds of papers every year.
I'm pleased with ACM's recent new options for publishing, which will serve to make publications much more accessible. However, we need to go further and begin to change our notions of peer review to make it more broad-based, speedy, and dynamic. And, we should consider adopting more flexible forms of publishing, using the technologies of text processing, revision management, global access, and social networks that our field has pioneered. I remember a meeting with the late Jim Gray, some years back, where he even then emphasized this need.
How do you convey to colleagues your sense that the role of computer science is expanding beyond traditional definitions to applications that span multiple disciplines?
The evidence is everywhere. Only 40 years ago, the field was more of an island unto itself. In those days, I suspect a very large percentage of us wrote programming languages and tools aimed only at ourselves! But, today, we cannot pick up a newspaper without reading about CS's actual or potential roles in education, healthcare, government, finance, etc.
So, I naturally advocate a cross-disciplinary approach: It fosters a broader cross-section of individuals wanting to enter the field. It allows us all to have more of the types of impact to which I previously alluded. And, as the field has become not just mathematical and engineering oriented, but also empirical, we need the data we get by applying CS everywhere. Finally, we contribute to the other disciplines, which, frankly, need us.
At Google, a cross-disciplinary approach allows our various research and development teams to draw upon a large ecosystem of knowledge, delivering innovation that perhaps would not be possible otherwise. Also, as in "Google's Hybrid Approach to Research," I advocate blurring the boundaries of R&D so we benefit from each other's skills and focus.
What advice would you give to budding technologists about the problem-solving potential of computer science?
Good time to ask as I've just been talking to many of Google's interns.
I begin my discussions with them by talking about "the absence of real limits." We will continue to build ever-more powerful computing engines. We've not begun to max out on the creative ideas we have to build intelligent systems. There are endless opportunities for beneficial impact to society. Then, I emphasize the breadth of the field, and the opportunity to have so many different roles, ranging from the invention of new algorithms and methodologies, to the conception of radically new products, to solving the true Grand Challenge problems.
I inform all the interns that I think the field will continue to be one of rapid change for the foreseeable future. While that will make it fun and valuable, it will also necessitate continuing education. All of us need to devote lots of energy to continuing education: the moment we stop learning, we become much less useful. It's exciting to see all the educational materials, and from diverse sources, now available, so that lifelong education can be a true continuum. (In addition to all the university MOOCs we hear so much about, even Google is offering them so that our users understand what our services do; e.g., advanced search, mapping, and YouTube production.)
Finally, I emphasize that they should always focus on impact. The real value of our individual achievement helps our society and compounds itself throughout our careers.
*∀ (upside down A) is called the universal quantifier, and means "for all values of x in the domain."