On Thursday, May 3, 2018 at 3:02:24 AM UTC+8, Xiongqi ZHANG wrote: > The judging system is supposed to be working like a black box test.
During a contest, that's ok. For practice there is no reason for it to be a black box. > Keep in mind that you can always come up with your own test data. You just > need a random test case generator and a naive but correct solution. Yes but that's a lot of extra work, and counterexamples can be hard to find. > If you have access to the data set that your implementation will fail, you > are likely just fixing the symptom, not the root cause I can't imagine why anybody would ever consider just fixing the symptom rather than the root cause, just to get it to work on a given test case instead of getting it to be actually correct. I would be tempted to say that such a person has some kind of mental problem. The whole point of testing with a data set is to find and fix bugs in the implementation (or in the logic). Anything else will be useless for the next contest. > Also, there are plenty of online judging systems that don't reveal their data > set (in practice mode too) Argumentum ad populum > Meanwhile, I agree that more complicated and non-trivial test case can be > provided as the sample input/output in practice mode (not the full one used > in a black box test though). If those test cases catch most bugs in the solutions, then that would be a good option. It's not so important to get the full real data sets. So we can agree on that :) -- You received this message because you are subscribed to the Google Groups "Google Code Jam" group. To unsubscribe from this group and stop receiving emails from it, send an email to google-code+unsubscr...@googlegroups.com. To post to this group, send email to google-code@googlegroups.com. To view this discussion on the web visit https://groups.google.com/d/msgid/google-code/8c1e7fb5-ac59-4f4b-ac67-77de70be2dcc%40googlegroups.com. For more options, visit https://groups.google.com/d/optout.
