I am a research student in the area of programming language/software engineering. I am claiming to provide an automated technique to solve a research problem in one of my papers. The prototype tool that I am creating for this paper is required to test some real-world libraries. I am using a symbolic execution tool in my research and it crashes on encountering native functions in some third-party libraries. To mitigate this problem, I provide my own model class that is basically a dummy class with non-native functions. Instead of using the original class which contained the native functions, I change the library to use this model class so that symbolic execution works better. Will this be called manual effort in my tool? Can't I call my technique an automated one now? Moreover, is it cheating?
1
-
1$\begingroup$ How about asking your advisor? $\endgroup$– Yuval FilmusOct 8, 2020 at 7:56
Add a comment
|
1 Answer
$\begingroup$
$\endgroup$
You can do that. I have seen other papers that generate dummy classes like that (e.g., Klee). How useful this will be may depend on how much effort it takes to write those dummy classes and how many are needed.
What matters is less the name (automated vs manual) and more that you describe clearly in your paper what its advantages and drawbacks are. Usually it is valuable to compare to prior papers. Can you demonstrate that your method is easier to apply than prior work? Maybe because it requires less manual work? If so, you may be in business.
