Resumen en español

Automated code generation is fostered by several software development methods. This generation is often supplied by well-known CASE (Computer-Aided Software Engineering) tools. However, automation is still so far and some CASE tools are complemented by non-standard modeling projects. In this paper, we conceptualize projects related to automated code generation, starting from discourse representations in either controlled or natural language, or in conceptual schemas. In this way, we present a graphical summary of crucial concepts related to this issue, by means of a state-of-the-art review. We conclude that automated code generation usually begins from solution-based representations of the problem instead of domain-based representations. Also, we summarize that these starting points are misunderstood by the client and this situation leads to poor validation in early stages of software development lifecycle.