Блог пользователя shen01

Автор shen01, история, 4 года назад, По-английски

With an increasing number of cases of cheating in Codeforces and Codechef contests, I am trying to set up a Multi-platform code plagiarism checker using MOSS tool developed by Stanford.

The plagiarism checker is working perfectly fine, now I just want to figure out a way to download all Accepted solutions of a SPECIFIC PROBLEM from all users to run the test on them. Can anyone help me with it?

  • Проголосовать: нравится
  • +18
  • Проголосовать: не нравится

»
4 года назад, # |
  Проголосовать: нравится -11 Проголосовать: не нравится

I am sorry but MOSS can be easily bypassed.

  • »
    »
    4 года назад, # ^ |
      Проголосовать: нравится +14 Проголосовать: не нравится

    Yup, I agree, but as you might have noticed, in most of the cheating that takes place in Codeforces contests, users either submit the same code or just change some variable names or indentation. And MOSS works pretty well in those cases.

    So it surely will help to reduce cheating.

»
4 года назад, # |
  Проголосовать: нравится 0 Проголосовать: не нравится

Codechef can already detect same code and variable name changes in the code

For Codeforces ,

It can detect exactly same codes and sometimes , if there are less variable changes ...

Makes sense , because most of the Div 3.A and Div2 A have a one line answer like max(a1 , a2) , n%2 or __gcd() , etc or a simple for loop

Does your MOSS tool do something different , becuz AFAIK , most of the online judges use MOSS

  • »
    »
    4 года назад, # ^ |
      Проголосовать: нравится +5 Проголосовать: не нравится

    Div3 A and Div2 A are not the ranks deciders in most of the cases. So won't run MOSS on them.

    Cheating in problems like Div2 C and onwards affects the overall standings mainly.

    What I have witnessed is that a group of people divide problems among them, share the code after pretests passed, and others do a variable change, and submit it. And as I said earlier, MOSS works in those cases

    • »
      »
      »
      4 года назад, # ^ |
        Проголосовать: нравится -8 Проголосовать: не нравится

      Div3 A and Div2 A are not the ranks deciders in most of the cases

      Well I totally disagree with that , Educational rounds ranks are decided on who solved the most questions and then time penalities , It doesn't give priority (or high scores ) to C or D ... , all problems have the same significance ..

      Keeping that aside , My main point was this

      Does your MOSS tool do something different , becuz AFAIK , most of the online judges use MOSS

      Because I don't think that codechef can have a better plagiarism detector ...

      • »
        »
        »
        »
        4 года назад, # ^ |
          Проголосовать: нравится +8 Проголосовать: не нравится

        Does your MOSS tool do something different , becuz AFAIK , most of the online judges use MOSS

        First of all, MOSS is not something which I have built. I was just trying to use it with the Contest submissions. I don't know what tools do Codechef and other online judges use.

        It's just an experiment from my side to check its efficiency and flaws of MOSS with such submissions.

        Also, I never planned to deploy it as a proper Plagiarism checker. Currently, it's for personal use only.

»
4 года назад, # |
  Проголосовать: нравится +11 Проголосовать: не нравится

Someday (today) cheaters will know they don't have to share codes, sharing ideas is enough.

  • »
    »
    4 года назад, # ^ |
      Проголосовать: нравится 0 Проголосовать: не нравится

    Yes, and I wonder if it's possible to build a plagiarism checker to counter that.

    But spreading ideas takes much more time and explanation (for complex problems) than spreading code.