Hello everyone, I just want share my idea for a simple bruteforce system to check for a repeat problem. What I've thought of won't prevent repeat problems from being written, but perhaps prevent them from being used in contest.
I propose that when a problem and code for the intended solution is written, all previous problem test cases are tested on it to see if anything is fully accepted with code that is written for the new problem. If such a problem is found, it can be manually reviewed for if it's just a coincidence or a repeat problem. This obviously won't detect every repeat problem due to different input requirements/formats, but I think it could help at least a little.
An example that would be submitted for manual review:
1804A - Lame King and 1452A - Robot Program while the problems aren't exactly the same, the same code could work for both.
The problem author and contest organisors could then decide wether they should replace the problem or if it's dissimilar enough to use the newly written problem.
The example problem I gave above is simply due to my limited experience and I'm sure people could find other harder problems that would be the same.
Whilst this is by far not a perfect solution, it could be a small step towards helping with repeat problems showing up in contest. Thanks for reading!
Side note: I know this is probably another useless blog and will never actually be used, but I thought I should share on the off chance it isn't a useless idea.