Morphological synthesis of exercises in the training system of programming

A part of teaching materials training system programming should include a large set of similar variants of tasks for writing programs for all the themes of the studied material. For a fixed set of tasks, the system loses its learning function as soon as the users executes all the tasks Therefore, the training system must be able to generate an option of the tasks "on the fly". In addition to the test tasks, the training system must comprise tasks for writing and editing code of different types: writing complete programs, writing a part of the program, fixing an error, refactoring a program code, writing a module text for the given code, etc. The generation of the variants of the tasks can be performed on the basis of the morphological synthesis. A typical task on programming on a certain subject should be a parameterized template with parameters. A specific version of the task is formed by the system by means of selecting parameter values. The combinations of the parameters determine the variety of the options. Based on the analysis of the specific tasks, three categories of the parameters: independent, dependent and nested are identified. There are two types of the values of the parameters: a set of string values and numeric values calculated by the system. When creating a particular option, the training system must first choose the values of the independent parameters. The values of the dependent parameters are chosen on the basis of the known values of independent parameters. And, for each value, which is associated with the nested parameters, the system must generate their values. To view the template in the base tasks the training system should have a domain specific language, such as xml-based. A specific task is shaped by textual substitution of the specific parameter value to the appropriate place in the body of the template in the same way as almost any macroprocessor does. To generate arithmetic expressions with standard features in tasks, a probabilistic algorithm that generates a grammar expression is developed.