摘要: UML modelling tools provide poor support for composite state machine code generation. Generated is typically complex and large, especially machines. Existing approaches either do not handle this case at all or it by flattening the into a simple one with combinatorial explosion of states, excessive generated code. This paper presents new approach that transforms an equivalent set machines before avoids leads to more concise scalable We implement our in Umple. report on study, comparing others terms size scalability.