The details of how timers work varies widely between operating systems and between variants of the same operating system (different “flavors” of UNIX, for example). IDL’s timer module is intended to provide a stable interface to the rest of IDL, and to isolate the non-portable code in one place.
Under UNIX, IDL’s timer module performs a more important function. UNIX processes contain a single timer that must be shared by the code in the process. When the timer fires, it raises the SIGALRM signal which must be caught and handled by the process. The IDL timer routines transparently multiplex this single timer to provide multiple virtual timers.
Under UNIX, IDL provides both blocking and non-blocking timers. Blocking timers put the calling process to sleep until they go off. Non-blocking timers are delivered asynchronously when they fire.
Under Microsoft Windows, only the blocking form of timer requests are supported.