Rambles around computer science

Diverting trains of thought, wasting precious time

Tue, 07 Apr 2009

The numbers game

I just ran across this truly excellent article by David Parnas about the folly of measuring success and impact by publication counts and citation counts. As you'd expect from the man, he says it all far better than I could, so I won't repeat anything. I couldn't let it go without blogging about it though. (I love the appropriateness of the one reference in the article, which is, very deliberately I'm sure, a negative citation.)

Early in one's career it's easy to feel insecure about one's competitive standing and, consequently, to end up participating to some extent in the various games he mentions. I think I'm fairly hard-line about “not playing the game” generally, but can't claim I haven't at least half-caved to one or two of the things he mentions (mainly the “submitting half-baked work” thing). So the article is quite helpful as a “sin list” to remind yourself why you're really doing research, and how it should work.

Of course I can't see things changing for the better any time soon on any of the counts he lists. But here's a thought: researchers typically want to work with like-minded individuals. Mismatched attitudes on this topic are easily detected---those who want tons of dubious publications will go and get them, achieving that flabbier CV that they always wanted. Employers who value this will immediately find the people they're looking for in this way. So in effect, attitudes to publicatinos become a filtering criterion by which mismatched pairings of researcher--institution or research--professor can be avoided. I may not have many publications, but if that just means I'm only likely to be employed by someone who understands the fraud of the game, and actually reads and likes my work, that suits me just fine. Of course I'm making the dangerous assumption that sufficiently many senior researchers actually have sensible views on the topic, so if I'm wrong, I may be out of a career....

[/research] permanent link contact


Powered by blosxom

validate this page