3 ms·

I don't agree with the parting thought, though. If you, as a PhD student, do make your project open source, document it and make it useful for others, and in so

I don't agree with the parting thought, though. If you, as a PhD student, do make your project open source, document it and make it useful for others, and in so doing, get a bunch of citations, you have an immediate career boost. How likely that is, is another question I am in a similar position as the author, some software I helped write on the side during my thesis is the only recognized remnant and actually got me my current job (tenured staff scientist)

mattkrause 7y ago
How do you address the issue that publications and citations are a one-and-done “reward”, while maintenance can go on forever. Do you publish a new paper every few years (“Marmaduke’s Modules 3.0”), so that you can show ‘productivity’? I know that this is a hack to fit software into the usual credit scheme that academia uses. However, being a software maintainer might be more like being a journal editor (another category that gets you some points for tenure/promotion), especially since both having ongoing responsibilities.
- Mathnerd314 7y ago
  Haskell has the Haskell Symposium: https://www.haskell.org/haskell-symposium/ https://www.haskell.org/haskell-symposium/ IDK how papers there count for tenure, it probably depends on the university.
- ken 7y ago
  Because the exchange rate isn't 1 release : 1 paper. When I worked as a programmer in academia, my lead professor said he'd had some folks from the computer science department over to look at what we were doing, and they commented that there were at least a couple of PhDs and 5 or 6 masters theses in there. I don't know if that's true (and I didn't stick around long enough to find out) but I know that fully organizing and documenting every bit that I'd done would take many times longer than writing it in the first place.
  - mattkrause 7y ago
    I'm not even sure there is a fixed exchange rate, but if there is, I think it's >> 1 release per paper[0]. I can definitely imagine writing a paper for the first major release. If the software solves an pressing scientific problem (e.g., replacing laborious experiments with accurate simulations) or shows off a new algorithm, the paper itself might be a big deal. Subsequent versions seem trickier. You could certainly publish big technical advances, like a parallel version of the core algorithm. I'm not sure what you do for more routine maintenance and incremental improvements (e.g., bug fixes, more import/export formats, better UI). There are specialized journals that might take these "reports", like the Journal of Open Source Software, but I'm not sure if a CV of JOSS papers would give you a viable career. [0] If I had to guess, I'd bet there were multiple PhDs in the topic you were working on. While the code base would be a good start, there's a lot more to a research degree than organizing and documenting some code.
nitwit005 7y ago
I'm curious what sort of software you wrote that people gave you a citation reliably? It often seems software fails to get mentioned unless it's central to the work. A few open source projects nudge users to cite them since it doesn't seem to happen naturally.

mattkrause 7y ago

How do you address the issue that publications and citations are a one-and-done “reward”, while maintenance can go on forever. Do you publish a new paper every few years (“Marmaduke’s Modules 3.0”), so that you can show ‘productivity’? I know that this is a hack to fit software into the usual credit scheme that academia uses. However, being a software maintainer might be more like being a journal editor (another category that gets you some points for tenure/promotion), especially since both having ongoing responsibilities.

Mathnerd314 7y ago
Haskell has the Haskell Symposium: https://www.haskell.org/haskell-symposium/ https://www.haskell.org/haskell-symposium/ IDK how papers there count for tenure, it probably depends on the university.
ken 7y ago
Because the exchange rate isn't 1 release : 1 paper. When I worked as a programmer in academia, my lead professor said he'd had some folks from the computer science department over to look at what we were doing, and they commented that there were at least a couple of PhDs and 5 or 6 masters theses in there. I don't know if that's true (and I didn't stick around long enough to find out) but I know that fully organizing and documenting every bit that I'd done would take many times longer than writing it in the first place.
- mattkrause 7y ago
  I'm not even sure there is a fixed exchange rate, but if there is, I think it's >> 1 release per paper[0]. I can definitely imagine writing a paper for the first major release. If the software solves an pressing scientific problem (e.g., replacing laborious experiments with accurate simulations) or shows off a new algorithm, the paper itself might be a big deal. Subsequent versions seem trickier. You could certainly publish big technical advances, like a parallel version of the core algorithm. I'm not sure what you do for more routine maintenance and incremental improvements (e.g., bug fixes, more import/export formats, better UI). There are specialized journals that might take these "reports", like the Journal of Open Source Software, but I'm not sure if a CV of JOSS papers would give you a viable career. [0] If I had to guess, I'd bet there were multiple PhDs in the topic you were working on. While the code base would be a good start, there's a lot more to a research degree than organizing and documenting some code.

nitwit005 7y ago

I'm curious what sort of software you wrote that people gave you a citation reliably? It often seems software fails to get mentioned unless it's central to the work. A few open source projects nudge users to cite them since it doesn't seem to happen naturally.