Compiling Techniques (CT)

Response to 2016/17 survey feedback for CT.

I have taken note of the feedback from students in this course survey, and have the following comments in response.

  • Course disorganised: It is true that we had some issues this year with the last assignment at the end of the course (writing a compiler pass in a real industry-grade compiler). This was was added following feedback from the previous year where students expressed the desire to learn about an actual compiler and we were probably not as prepared as we could have been. Since we will be running this for the second time this year, I am confident that this should go more smoothly this time.
  • Test suite source code hidden: Many students complain that we did not release the source code of the test used to mark the compiler. This is actually on purpose to emulate real-world condition where software developer should be writing their own tests. Our automatic test suite gives a daily score and this should be seen as an indication of how well the compiler behaves with respect to the specification.
  • Negative scores: Having negative scores is unfortunately necessary since for part 1 and 2, the goal of the compiler is to determine whether a program is valid or invalid. Without negative scoring, a compiler that would randomly generate an answer (whether a program is valid or invalid) would on average score 50%.
  • Feedback late: All the marks have always been returned on time to the ITO office. The score on the hidden tests were also pushed into each student repository as soon as the tests were run. We will make this clearer this year in case there is any doubt on how to access this information.

Christophe Dubach, September 2017