On 20 Jan 2011, at 01:36, Asst. Prof. Dmitrii (Dima) Pasechnik wrote:
>
> Mostly 1), 2), and 5) --- namely, we would like to examine the ability
> to write elementary (and working!) pieces
> of computer code. Doing this on paper is a pain.
Every Computer Science department I know uses homework for this rather than
exams. Exams might ask for a few lines of code, usually not worrying too much
about perfect syntax, or, more commonly, present code fragments and ask
questions about what they do -- this can be quite subtle and test understanding
well. Asking students to write and debug code in exam conditions is generally
too much grief.
It is possible to run "exam-quality" lab sessions -- we call them "skills
tests", but it's a lot of work. You need to make the instructions very clear
indeed, provide extra computers (and some means of recovering work done so far)
for students whose machines fail or freeze or whatever, turn off internet
access or have enough invigilators that you can reliably spot anyone using
email or chat, and have different version of the test for each "lab-full" of
students and then make sure the marks are comparable.
Of course homework raises cheating issues, but it's pretty easy to spot two
students whose code only differs in variable names and comments.
Steve
_______________________________________________
Forum mailing list
Forum@mail.gap-system.org
http://mail.gap-system.org/mailman/listinfo/forum
