When working on the theory of constraint satisfaction problems, one usually restricts to a finite set of equal variables, which can take values from a finite domain (see chapter 1, 1.1 in [1]). For many scheduling problems, this is not appropriate. For example: a schedule for the camera on a satellite shall be created. The visibilities of the targets usually won’t appear as a multiple of a considerable large time unit. Besides the tasks will often not be equal but some of the tasks can build groups, which shall be scheduled as a unit (e.g. the uplink, which sends the command to the satellite, the datatake, i.e. the taking of the picture and the downlink, which sends the picture to the ground station belong together). The aim of this article is, to introduce an object language, which allows to model all sorts of constraints in an efficient way and which also enables the user, to define a structure, which can help the algorithm to be more efficient.