Models of estimation in instruction programming system

The article deals with the methods of students’ knowledge level estimation in automated instruction systems designed for study of programming sciences. Three models are defined: a domain model, a task model and a student’s knowledge model. Student’s knowledge level is determined by computable supervision — test execution estimation, which is calculated as a sum of test execution accuracy and code quality estimation. The last one is formed as divergence estimation between the etalon program model and the model generated in accordance with student actions.