Competitions in informatics are usually synonyms for algorithmic programming contests. Many of these competitions use automatic grading of the contestants’ solutions. This is done by running them on batches of input data and testing the correctness of the output. In this paper we will introduce a newly developed system called “MENDO”, which is used by the Macedonian Computer Society in the organization of national informatics competitions. We will outline the main modules every grading system should support (sandbox, grader, controller, auxiliary modules) and present how each one of them is implemented in our system. We will show that MENDO is comprehensive and superior in a number of functionalities as compared to the other presently existing systems.
programming contests, algorithmic problems, grading systems, automatic grading, task evaluation, contest management system