Re: What's Real-Time? (was Re: Widespread C++ Competency Gap?)

[przemek@rrdjazz.nist.gov (Przemek Klosowski)]

In article <3e5t1e$2ad@oghma.synapse.net> rjc@oghma.synapse.net (Robert J Carter) writes:

   >large 'hard' systems in the traditional sense are simply too
   >wasteful. An example is the telephone system. It is not 'hard real
   >time': the cost would be just too big. Instead, they figured out how
   >to manage available resources to have acceptably low failure rates.

   That's fine as far as it goes, but having done some work myself with 
   reactor systems (I saw your sig), what is an acceptably low failure rate
   for the emergency shutdown system of a reactor?

Of course the only acceptable failure rate for reactor emergency
shutdown is zero.  I was talking of failure in the sense of some tasks
not meeting their deadline: e.g. if we have time, we shut down the
pumps gradually, but if we are behind deadlines, we just scram them.

The traditional ('bigoted', as the Dr. Dobbs article had it) real time
has a discontinuous 'utility' function: one if the task finishes
before deadline, minus infinity otherwise. The article claims that a more
realistic approximation is a function like:


       utility to the system
                  ^
                  |
        ----------|1
                  |\
                  | \
   ---------------|--\--------> tardiness ( = time of completion - deadline)
                 0|   \
                  |    |

		
BTW, I have nothing to do with the reactor design and operation: I am
a solid state physicist, doing neutron scattering.

--
			przemek klosowski (przemek@rrdstrad.nist.gov)
			Reactor Division (bldg. 235), E111
			National Institute of Standards and Technology
			Gaithersburg, MD 20899,      USA

			(301) 975 6249