Many detectors are constructed from modules, with "dead pixels" in the gaps (thinking of Eiger and Pilatus in particular here) - we have the structure of these detectors encoded all over the place for defining e.g. the mask regions and such like.
There are however good arguments for properly encoding the structure e.g. to allow calculations to be split across independent modules - without necessarily splitting the data array. Suggestion is along the lines of defining some non-overlapping regions of the panel which corresponds to individual modules, probably as pixel ranges since everything else can be derived from this.
Many detectors are constructed from modules, with "dead pixels" in the gaps (thinking of Eiger and Pilatus in particular here) - we have the structure of these detectors encoded all over the place for defining e.g. the mask regions and such like.
There are however good arguments for properly encoding the structure e.g. to allow calculations to be split across independent modules - without necessarily splitting the data array. Suggestion is along the lines of defining some non-overlapping regions of the panel which corresponds to individual modules, probably as pixel ranges since everything else can be derived from this.