Model Transformation Language MOLA Audris Kalnins, Janis Barzdins, Edgars Celms University of Latvia, IMCS, 29 Raina boulevard, Riga, Latvia {Audris.Kalnins, Janis.Barzdins, Edgars.Celms}@mii.lu.lv Abstract. The paper describes a new graphical model transformation language MOLA. The basic idea of MOLA is to merge traditional structured program- ming as a control structure with pattern-based transformation rules. The key language element is a graphical loop concept. The main goal of MOLA is to de- scribe model transformations in a natural and easy readable way. 1 Introduction The Model Driven Architecture (MDA) initiative treats models as proper artifacts during software development process and model-to-model transformations as a proper part of this process. Therefore there is a growing need for model transformation lan- guages and tools that would be highly acceptable by users. Though model transforma- tions would be built by a relatively small community of advanced users, the prerequi- site for broad acceptance of transformations by system developers is their easy read- ability and customizability. Model transformation languages to a great degree are a new type of languages when compared to design and programming languages. The only sound assumption here is that all models in the MDA process (either UML-based models or other) should be based on metamodels conforming to MOF 2.0 standards. The need for standardization in the area of model transformation languages led to the MOF 2.0 Query/Views/Transformations (QVT) request for Proposals (RFP)[1] from OMG. To a great degree the success of the MDA initiative and of QVT in particular will depend on the availability of a concrete syntax for model-to-model transformations that is able to express non-trivial transformations in a clear and compact format that would be useful for industrial production of business software [2]. QVT submissions by several consortiums have been made [3, 4, 5], but all of them are far from a final version of a model transformation language. Currently the most advanced proposal seems to be [3]. Several serious proposals for transformation lan- guages have been provided outside the OMG activities. The most interesting and complete of them seem to be UMLX [6] and GReAT [7]. According to our view, and many others [2], model transformations should be defined graphically, but should combine the graphical form with text where appropriate. Graphical forms of transformations have the advantage of being able to represent mappings between patterns in source and target models in a direct way. This is the motivation behind visual languages such as UMLX, GReAT and the others proposed