mouse@Holo.Rodents.Montreal.QC.CA (der Mouse) writes:
> The basic problem here seems to me to be that different programs want
> to do different things.  In particular, some processes want to sleep
> until a specific time; others want to sleep for a specific interval.
> Unfortunately, the current itimer interface cannot satisfy both of them
> [...]

Agreed.

Since API changes have been mentioned in this thread, let me just note
that POSIX.1b timers can be set with either absolute or relative
triggers.  

It seems to me that instead of implementing some new API, we could
solve this problem by implementing the POSIX timer interface.  The 
traditional API can be provided by userland wrappers that call the
new syscalls.  I already do this with POSIX clocks code I've 
written.

	--jtc