better scheduler with correct sleep times

[greg]

sokol wrote:
> Also, the
> scheduler runs inside a loop. How do you suppose to
> run other code while the loop is executing?

The sleep function could be doing a select with a
timeout on some other source of events, such as a
socket or a gui input event stream. One possible
response to such an event is to schedule another
event. There's not so much need for that nowadays,
since most gui libraries provide a way of scheduling
timed events as part of their built-in event loop,
but you might want to use something like this in
a server that deals with network connections.

Another possible use is discrete-event simulation,
where the "sleep" function doesn't physically sleep
but just advances a simulated time, and all events
(other than the first one that starts everything off)
are scheduled by callbacks for other events.

So while its uses are rather specialized, I wouldn't
say it's useless. The main problem is that its nature
needs to be much more clearly spelled out in the
docs -- it's something of an attractive nuisance the
way it is.

-- 
Greg