Automatic translation from pseudocode to source code: A conceptual-metamodel approach

Pseudocode is commonly used to teach algorithms for computer science students in an introductory programming course. When an automatic translation tool is available, students can translate their pseudocode to source code and then execute the resulted source code to check whether their algorithm is correct. However, a large number of translation tools are needed to accomodate courses in different situations: (i) different natural languages to define pseudocode, (ii) different programming languages to implement pseudocode, and (iii) different templates to write pseudocode. Developing many tools from scratch requires large effort. This paper presents an approach for developing many translation tools in an efficient way. The approach uses an intermediate model to decouple pseudocode translation from source code translation. The approach is based on a conceptual metamodel that has to be realised by pseudocode and intermediate models. This approach allows the development of translation tools by reusing parts of already-developed translation tools.