Abstract

Favre, Jean-Marie. (2005). Megamodelling and Etymology.

Is a model of a model, a metamodel? Is the relational model a metamodel? Is it a model? What is a component metamodel? Is it a model of a component model? The word MODEL is subject to a lot of debates in Model Driven Engineering. Add the notion of metamodel on top of it and you will just enter what some people call the Meta-muddle. Recently megamodels have been proposed to avoid the meta-muddle. This approach is very promising but it does not solve however the primary problem. That is, even a simple use of the word Model could lead to misunderstanding and confusion. This paper tackles this problem from its very source: the polysemic nature of the word MODEL. The evolution and semantic variations of the word MODEL are modelled from many different perspec- tives. This papers tells how the prefix MED in indo-european has lead, five millenniums after, to the acronym MDE, and this via the word MODEL. Based on an extensive study of encyclopedias, dic- tionaries, thesauri, and etymological sources, it is shown that the many senses of the word MODEL can be clustered into four groups, namely model-as-representation, model-as-example, model-as-type, and model-as-mold. All these groups are fundamental to understand the real nature of Model Driven Engineering. Megamodels and Etymology are indeed keys to avoid the Meta-muddle.

  Presentation

Presentation [EXT]