2

I have a month to finish an undergraduate research project required to receive my diploma and I am facing the situation when an undergraduate is too low in the academic scale to fight his/her professor, a very stubborn one indeed. Just to say that in a previous situation I warned him/her about certain approach that I knew was not going to work (because I checked) and, besides being scolded, was obliged to loose an entire month of my project doing menial things.

Yes, I made mistakes too. This is my first research project in his/her field and some scripts needed corrections. I don't want to sound that harsh. The professor went discontented and very stressed for a couple of weeks due to finals and similar stuff and, during that time, the amount of data exploded. Furthermore, He/She is an excellent Human, willing to support you in every possible situation, but, when it comes to work, I have been feeling like Sisyphus, pushing a stone that goes nowhere for most of the project and then obliged to fix up everything that went wrong (and made responsible of it).

I put sweat and tears in the project expecting, after all the ordeal, something good will come out, but this definitely won't happen. The project involves funding and my professor wants the most accurate results, but I have to think about myself and my graduation too. This is Computer Science and, although we have lots of machine power, there are always amounts of data beyond our capacities. Did anyone in computer science had a similar problem with his/her advisor? What did you do?

  • 1
    What does your professor say? What is his/her point of view on the topic? Does he/she acknowledge that the problem is not solvable? – Federico Poloni Jun 15 '15 at 6:14
  • 2
    Right now, I have to admit I'm having a hard time understanding your question. Could you focus on specifying that more precisely? – aeismail Jun 15 '15 at 15:36
  • 6
    I have been feeling like Sisyphus, pushing a stone that goes nowhere for most of the project — So, real research, then. – JeffE Jun 15 '15 at 16:03
  • @FedericoPoloni I think he/she is not aware of this fact yet. Although, I tried to make it clear once, but the professor said, in summary, "run the process, your are wrong..." That was past friday. I checked the process today and is still in very beginning. – CSprog Jun 15 '15 at 16:55
  • This might suck to hear, but I've found that if you bang your head against something long enough you'll find some kind of solution, even if its a bizarre trade off of one thing for another. Case in point, I've written a real time raytracer in JavaScript / html5 (not webgl!). Basically I got "dirty rects" working for raytracing so the number of rays cast were very low. I hope you can find an imaginative solution to your problem (: – Alan Wolfe Jun 15 '15 at 17:52
4

So, if I understood you correctly, your professor wants you to run some time consuming program on the data, and it will not be finished before the deadline for your graduation, right?

If that is the case, you can just include in your report the subset of your data that is finished, indicating that the rest is being processed. Try to make that sample as representative of the whole, if that makes sense in your case (so, if your data goes from x=0 to x=100, try to get evenly spaced values of x, not just from x=0 to x=1.5). After all, what you need to show for an undergrad project is that you know how to do research, but not necessarily get results all the way.

Another thing is that if in four days only 1% has been processed, it would take a full year to process the whole dataset. You seem to need* either more computer power, or rethink the strategy; but now you have some hard data to bring up to your professor. Then, it is up to him to decide if he can get access to more CPU muscle, or if a compromise in the quality is acceptable.

  • Check also your cluster occupancy. If you are sharing it with other students with the same deadline, it is possible that it very crowded; more resources may be available after the deadline has passed.

Your Answer

By clicking “Post Your Answer”, you agree to our terms of service, privacy policy and cookie policy

Not the answer you're looking for? Browse other questions tagged or ask your own question.