Abstract: A model of a geometrical object comprising: bits in a first data structure encoded to identify a first set of elements from a collection of representations of geometrical shapes for a collection of elements and to represent positions of the elements by means of integer co-ordinates in a first system of co-ordinates; bits in a second data structure encoded to identify a second set of elements from said collection and to represent positions of the elements by means of integer co-ordinates in a second system of co-ordinates; bits in a third data structure encoded to represent a spatial transformation of the second system of co-ordinates relatively to the first system of co-ordinates. Thereby it is possible to create a very compact representation of a model of a geometrical object. The model—prima facia—seems to be constrained very hard, but is in fact a very flexible and easy to access model.