Dynamic modelling of lymphocyte behaviour has primarily been based on populations based differential equations or on cellular agents moving in space, and interacting each other. The final steps of this modelling effort are expressed in a code written in a programming language. On account of the complete lack of standardization of the different steps to proceed, we have to deplore poor communication and sharing between experimentalists, theoreticians and programmers. The adoption of diagrammatic visual computer language should however greatly help the immunologists to better communicate, to more easily identify the models similarities and facilitate the reuse and extension of existing software models. Since immunologists often conceptualize the dynamical evolution of immune systems in terms of state-transitions of biological objects, we promote the use of Unified Modelling Language (UML) state-transition diagram. To demonstrate the feasibility of this approach, we present a UML refactoring of two published models on thymocyte differentiation. Originally built with different modelling strategies a mathematical ODE-based model and a cellular automata model, the two models are now in the same visual formalism and can be compared.