1. 3

  2. 3

    Good comment. Some practical caveats, based on some experience collaborating with academics in CS (and related fields like HCI) while not being an academic myself:

    1. Most labs / research groups have their own particular ways of doing things, procedures, processes, etc. (resulting from anything from the PI’s / lab head’s idiosyncracies to institutional inertia to constraints imposed by funding sources). It can be a bit tricky to adapt to that.

    2. If you’re collaborating remotely, with a team consisting of multiple members of a research lab / group, remember that those people probably see each other every day, so it can be a bit tricky to stay “in the loop” in terms of being on the same page, having the same concepts “loaded” into your head when thinking / discussing about the project, etc.

    3. The pace of projects done by academic researchers is not the same as the pace of projects “out in the world”. Certain other practices (such as anything having to do with development methologies, to take an easy example) may differ / be hard to apply.

    4. Success criteria may not be what you’re used to. (“A working product” vs. “a publishable paper” is a simplistic but roughly accurate way of putting it; differences in resulting incentive structures are also relevant)

    Still, such collaborations can be valuable learning experiences, and getting to take part in research of cool new stuff is very exciting!

    1. 4

      As a CS PhD student-cum-postdoc (so not exactly an academic but not exactly not). I very much endorse (4). If you’re asking an academic to do something, particularly if it’s something time-consuming or difficult like engineering work or some sort of community intervention, then your value proposition for them has to include a way in which this leads to an interesting publication.

      The problem, for anyone not currently up-to-date in a field, is often just understanding what might be interesting to an expert. That’s not necessarily an issue – your prospective collaborator no doubt has several possible solutions lying around waiting for a suitable problem, so if your project is flexible enough to accomodate this it could well benefit from experimental input – if you’re willing to suffer quietly while the bugs are ironed out. You provide data or other resources, and benefit from an erratic consultant.

      If on the other hand you have an idea of your own you want to try out, in collaboration, you might have to do the work to demonstrate that you’re not a crank, that you have at least some idea of the relevant academic background to what you’re attempting, and you might have to prepare yourself to be told that it’s been done already, or that it’s not novel enough.

    Recent Comments