For PHENIX, the central way for software to get at event data is the Event Handling Library, which is part of the ONCS software system. It is used to take care of almost every aspect of the raw data handling, such as getting access to the event data from the various sources of such data, to give you information about the events read, to pass them on to another destination, or to allow you to get at the contents in a simple way.

Fundamentally, the system consists of two different parts: a library, which comes with a well-designed API to allow to be used by your program or by ROOT, and a set of ready-made tools designed to give you information about the data stream in various ways.

Both parts contain a rich set of tools and functions, and make heavy use of the standard C++ design patterns such as factories, proxy objects, and so on. The goal is fast execution and an easy-to-use API.

If nothing else, you should probably read the "Events and Packets" chapter and about the dlist, ddump, and dpipe utilities.