3 ms·

Why do good software people leave academia?

5 points by williamstein 10y ago

PaulHoule 10y ago
When I was in grad school (physics) at Cornell I was making Java Applets for physics and engineering education, back when Java was in beta. I got invited to a CS conference in Syracuse on "Java in Scientific Computing" and did a live demo of the applets and was told I was very brave to do so -- all the actual CS people had nothing but screenshots. In academic CS the product is papers and conference talks, not software, so usually CS people are good at fleshing out ideas well enough to talk about them, but they aren't good at the fit and finish work that it takes to make real products you can hand off to other people in either industry or the academy. Academics in just about every other field (ex. economics, physics, biology, sometimes even the humanities) also develop software with the goal of actually developing software (to either do the work in some other field you are going to write papers about or to help in education,...) but they run into the problem that software development is not considered a core activity into their discipline.
- bachmeier 10y ago
  > Academics in just about every other field (ex. economics, physics, biology, sometimes even the humanities) also develop software with the goal of actually developing software (to either do the work in some other field you are going to write papers about or to help in education,...) but they run into the problem that software development is not considered a core activity into their discipline. That's right. I'm an economist, and for the most part, there is no incentive to write and share software. That's something that makes no sense to me at all, because the biggest constraint when doing research in some areas is the insane amount of time it takes to write computer implementations of the necessary methods. In general, writing software is viewed as "not research" because most economists don't understand what is involved. This is by no means an issue of marginal importance. Take this quote from Michael Keane[1]: "structural econometric work is just very hard to do, simply in terms of the amount of labor involved. It often takes several years to write just one good paper in this genre, and this poses a daunting prospect or the young assistant professor seeking tenure, or the graduate student seeking to finish a dissertation. Using 'simple' methods, one can write papers quicker." Given the choice between writing papers that carry the overhead of doing lots of programming for no benefit and writing papers that require no programming, most economists have chosen the latter. There's a real cost to the refusal to value software development in economics. [1] https://editorialexpress.com/jrust/research/JE_Keynote_7.pdf https://editorialexpress.com/jrust/research/JE_Keynote_7.pdf (page 41)

5 points by williamstein 10y ago

When I was in grad school (physics) at Cornell I was making Java Applets for physics and engineering education, back when Java was in beta. I got invited to a CS conference in Syracuse on "Java in Scientific Computing" and did a live demo of the applets and was told I was very brave to do so -- all the actual CS people had nothing but screenshots. In academic CS the product is papers and conference talks, not software, so usually CS people are good at fleshing out ideas well enough to talk about them, but they aren't good at the fit and finish work that it takes to make real products you can hand off to other people in either industry or the academy. Academics in just about every other field (ex. economics, physics, biology, sometimes even the humanities) also develop software with the goal of actually developing software (to either do the work in some other field you are going to write papers about or to help in education,...) but they run into the problem that software development is not considered a core activity into their discipline.

bachmeier 10y ago
> Academics in just about every other field (ex. economics, physics, biology, sometimes even the humanities) also develop software with the goal of actually developing software (to either do the work in some other field you are going to write papers about or to help in education,...) but they run into the problem that software development is not considered a core activity into their discipline. That's right. I'm an economist, and for the most part, there is no incentive to write and share software. That's something that makes no sense to me at all, because the biggest constraint when doing research in some areas is the insane amount of time it takes to write computer implementations of the necessary methods. In general, writing software is viewed as "not research" because most economists don't understand what is involved. This is by no means an issue of marginal importance. Take this quote from Michael Keane[1]: "structural econometric work is just very hard to do, simply in terms of the amount of labor involved. It often takes several years to write just one good paper in this genre, and this poses a daunting prospect or the young assistant professor seeking tenure, or the graduate student seeking to finish a dissertation. Using 'simple' methods, one can write papers quicker." Given the choice between writing papers that carry the overhead of doing lots of programming for no benefit and writing papers that require no programming, most economists have chosen the latter. There's a real cost to the refusal to value software development in economics. [1] https://editorialexpress.com/jrust/research/JE_Keynote_7.pdf https://editorialexpress.com/jrust/research/JE_Keynote_7.pdf (page 41)