Short background: I started my PhD last November and now I am helping a master student writing his MSc thesis. I still have problems when I need to formulate a concise research question and even, at least in workshop papers, some publications fail in formulating clear research questions as well.

The topic the CS student writes about is the following. There is this work by Bracha on pluggable / optional type systems, e.g. for scripting languages. The student wants to solve the same problem that these 'pluggable typesystem' solves. But he is using dependent-type theory, i.e., to check that values in the scripting language are valid according to a given type. This can solve (or better solve) problems related to scripting languages (e.g. security problems in web programming; because everything is basically a string in scripting languages).

I find it hard to come up with a concise research question. (And possibly also with a method to evaluate the approach).

So, my question is:

  • Are there any references that can help me to formulate valid research questions?

The closest reference I have found is this mini-tutorial by Mary Shaw.

From Germany there is also a Memorandum that is interesting, but the focus is only on information systems research and not CS. (I can't find a link to the long version in either English or German yet.)

  • You mean something like this? (What does your advisor have to say about that? I guess he shifted this supervision on you?) – Raphael Aug 29 '12 at 20:51
  • This paper is advice on how to properly referee/review a paper. In turn, it shows how to properly write a good research paper. – Nicholas Mancuso Aug 30 '12 at 0:51
  • This seems at least a little area-specific, especially between more "systems" and more "theory" oriented parts of CS. – Suresh Aug 30 '12 at 8:33
  • 1
    Another thing: look at (important) paperes of the field. They usually contain "open questions" or "future research". Try to figure out what the big players thought interesting but have not done yet, and start from there. – Raphael Aug 30 '12 at 21:27
  • 6
    Often the correct problem formulation only becomes clear after you have the solution in hand. It's relatively rare to actually solve the precise problem that you set out in advance to solve. (Or maybe that's just me.) – JeffE Sep 1 '12 at 15:46

Have you tried the following:

Turabian, Kate L. A Manual for Writers. Chicago: University of Chicago, 1987. Print.

It's a very nice book and is applicable to CS (unlike many other research methodology books).

| improve this answer | |
  1. Formulating good research problems comes from experience. If you understand an area well, you will see the fundamental concepts. From there, your problem statement is simply an articulation of these concepts. For a beginner, this is very hard. Advice: try to cut down the problems into sub-problems. Some will be easy, and some will be 'hard', and solving them means the rests become easy or open new doors. So construct a work breakdown structure (WBS).
  2. Try to formalize the problem; i.e., derive a mathematical model for the problem at hand. This will help focus your mind on key concepts or variables. For example, for an optimization problem, you might decide on an objective function, and after that determine all relevant constraints. This will be an iterative process.
  3. Start with a toy example, with as many assumptions that you need to make it a toy example. Then slowly generalize and once you have enough intuition, then formulate the key problem to be solved.
| improve this answer | |

Problem formulation is crucial phase in the research process. It start from; what is known at certain point of time and what is gape/defect/uncertainty/challenge/weakness in existing knowledge/system/model/solution or answer to a quest. Also Problem formulation starts from thrust/quest/hunger for new knowledge/knowing to unknown or need for exploration/extension/enhancement of the existing knowledge at a point of time. Finally Problem formulation precisely needs to define and draft the statement of the weakness in aforementioned knowledge using crystal clear words with preferably directional hypothesis. In statement it should cover 1)what is required?, 2) desired inputs and outputs with desired features and inter-linkage.

| improve this answer | |
  • 2
    I reviewed this answer because someone raised a Low Quality Posts flag. I must admit I failed to see why it's bad. It's not a spam nor abusive. I could not find any source it came from, i.e. it might not be plagiarized from somewhere. Would the downvoters explain why it's bad? Thanks. – scaaahu Dec 11 '17 at 3:06
  • @scaaahu I was the VLQ, but I didn't vote it down. I did it because I found it hard to understand. (Btw, the flag became "disputed".) Probably the downvoters had the same reason. Now I think maybe I was too eager to flag it, my habit is nicer in general. – peterh - Reinstate Monica Dec 11 '17 at 7:34
  • @peterh Thanks for replying. I was afraid I was missing something so I asked the question. Now, I understand. Yes, I agree this post was hard to understand. But, please trust me, I have reviewed thousands of posts and many of them are much more incomprehensible than this one. Sorry about the disputed flag. – scaaahu Dec 11 '17 at 7:52

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.