I was part of the NSERC Computer Science Grant Selection Committee again this year. Like last year, we met in Ottawa for a week of grueling but interesting reviews of grant applications on computer science research from all over Canada. There are two CS committees (330 and 331) but they are in the process of being merged and this year I was part of the review committee for a diverse set of projects. In some cases, there were proposals from people whose work I was familiar with, but in most of those cases I was not part of the processes, and often not even allowed in the room, due to both the vagaries of the semi-random selection process combined with the stringent conflict-of-interest guidelines.
For several decades the core of the scientific granting system has been the Discovery Grant (formerly the Operating Grant) that is awarded in a per-investigator basis. Most front-line active science researchers in Canada have one of these, and the modest funding levels were determined by annually making small adjustments to the researchers previous grant. For active performers, this meant a gradually increasing funding level over the course of one's career. All in all this system worked extremely well and, despite limited amounts of funding, it provided stable effective funding for Canadian research and training. In the last decade or so, however, limitations on the amount of available research funding have prevented a few researchers from "ramping up" as fast the the organization would have liked. As a result, a drastic new process was instituted last year in several areas (despite concerns from many of the GSC members), and came this year to Computer Science.
As of this year, award levels have been largely decoupled from past funding and, in fact, the GSC members vote on scores for the grants without really knowing in advance either who will get funding, what scores assure funding, or what the final cutoff will be. That is, the GSC awards various scores, but the translation of these scores into actual dollars is only determined later. Apparently this new system produces largely acceptable results, but as a member of the GSC is was quite disconcerting to be evaluating a proposal without being able to explicitly vote that it should be funded. Also, although past performance is going to be directly coupled to funding, there is, in principle, the potential for less stability in the funding rates. This will certainly cause trepidation on the part of researchers (and could thus lead to very conservative behavior the year before a renewal is requested). All in all, I will be very interested in how well this works and what changes in funding result. I suspect some strong junior researchers may experience big increases in funding while some more senior researchers may get an unpleasant surprise.
