In 2008 the ALMA project decided to put in two phases the development of its data model, the versions 1 and 2. The version 1 is currently used by ALMA and the EVLA. The definition of version 2 is elaborated in a broader context to allow to support a multiplicity of concepts of (radio-)telescopes. Covering two use-cases of very different concepts ALMA and EMBRACE (a prototype of a multi-beam aperture phased array in the design study for SKA), the development of the version 2 leads to a generic meta-model which should be adequate for any telescope. To keep this model simple it is necessary to go a step beyond the standard relational data model. This presentation will begin from the question of what is a data model. I’ll explain how I proceed to give to any component in the model, whatever its level in the global structure, a very concise meaning. The approach has mathematical foundations that I’ll briefly refer to. It leads to very robust (type-safe), efficient implementations and expressive applications.
Dans la même rubrique :