Publish and/or perish; proposals and reports; research, teaching and service: these are the "death and taxes" equivalents for life in major research universities. Success -- or at least promotion and tenure – is normally measured by how well one plays the academic game, publishing frequently in the "right" places, placing students in the "right" institutions and serving on the "right" committees. All too often, however, we tend to focus on these indicators rather than the presence or absence of impact from our activities.
Many of us know faculty who have published hundreds of papers, but who have not demonstrably changed the nature or direction of their specialty. Conversely, many of us can readily identify individuals who have transformed the nature of computing, but who have few – if any – of the standard beans that are counted when quantifying traditional research success.
As a department head, my guidance to young assistant professors was always simple: do something interesting and important, make a difference that is recognized, and the artifacts will come. The artifacts are just that -- things that accrue from impact, not objectives to be sought.
I was reflecting on all of this as I tracked the responses to Moshe Vardi's CACM editorial, Conferences vs. Journals in Computing Research, and Lance Fortnow's CACM article, Time for Computer Science to Grow Up. Before these two articles appeared, I had already discussed these same concerns about computing's shift from journal to conference publications in my March 2009 CRA column: Publishing Quarks: Considering Our Culture, where I bemoaned the rise of the minimum publishable unit and called for a reinvigoration of journals.
It is somewhat ironic that many aspects of our computing research culture are now so often conservative, especially when research itself is defined by an insatiable curiosity about the unknown and unexplained. If you have any doubt, try to secure research funding for a proposal that is controversial or try to move a curriculum revision rapidly through the academic approval process! (To be sure, there are exceptions and there are good reasons for many of these checks and balances, but the process is sometimes excruciatingly slow and often incremental.)
It is worth remembering that our field grew from the confluence of mathematics, physics, electrical engineering, psychology, information theory and management and a host of other disciplines. The founders were radicals and iconoclasts, eschewing the extant culture of their disciplines because this new notion of computing was so pregnant with possibilities that could transform society and the nature of discovery. In short, it was the most exciting thing they could imagine doing, even if it was not the "safe" thing to do. More to the point, they were not focused on getting multiple papers in this year's specialty conferences.
Remember, it's the impact that matters, not the artifacts.