The binary structure used by YODA for representing physical events is shown in 3.1. There are two main classes of eventstructures: containerevents and subevents. Both share the same header information: a 4 byte unsigned integer giving the total length (in bytes) of this containerevent / subevent (excluding this header), and a type/subtype identification (each an unsigned 2 byte-value), which is used to classify this event. This header is then followed by data, which meaning is given by the type/subtype information field. The important difference between container- and subevents is the fact that containerevents contain data + subevents, while subevents carry only their data.
Not all defined types/subtypes are used, but the user can (in principle) use all defined types or add new types/subtypes according to his or her needs. Both types and subtypes are actually 16-bit values (shorts), but are given in a symbolic notation here for readability.
Type and subtype information is used to label the data according to its meaning. In YODA's selector database, this information is needed to identify the data and give it its name, under which it it accessible inside the user tasks.
