Information specific to a particular instrument is not so easy to categorize. There is some information connected with individual passbands that can naturally be stored in tabular form: the codes used for filters, and the names of the bands, for example. But each particular type of detector has a different set of characteristics, which in turn require different sets of supplementary data. Furthermore, there are distinct modes of operation peculiar to each type: we need dead-time information about photomultipliers only when they are used as pulse-generators, not in DC photometry or charge integration.