🚀 go-pugleaf

-----BEGIN PGP SIGNED MESSAGE-----

In article <41i03t$q8u@sundog.tiac.net>,
Jim Williams  <williams@tiac.net> wrote:

>This does not seem to be a sufficient definition to me.  Any compute
>bound job would quickly turn such a system into a Unitasking scheme.
>A multitasking system must run mutiple tasks.  If one task can prevent
>the system from running multiple tasks then it is not a multitasking
>system -- once the task has taken control over the system the system
>has become a unitasking system.

it seems that you are not willing to admit as multitasking any system
which permits starvation.  it is rumored than when MIT took down their
IBM 7094 batch system in 1973, the found a job that had been waiting
to execute since 1967.  this system, none the less, was
multiprocessing.

when a process is running over a uniprocessor, it is the only program
running--until it terminates, relinqishes, or is preempted.  does this
make the system "unitasking"?  no.  eventually the process will
terminate or relinquish (or will be terminated), and eventually
another process will run.  the goal of multiprogramming
(i.e. multitasking) is to keep the CPU maximally utilized.  whether
this is accomplished by allowing one process all the ticks, or all of
the processes one tick each, is irrelevant, as long as all the ticks
are utilized.

you must realize that i may be approaching this from a broader
viewpoint than you are.

-----BEGIN PGP SIGNATURE-----
Version: 2.6.2

iQCVAwUBMDz8bBbgfSHT/piVAQHgWQP9EmhJH6kNHR9p4pHuFrcKrnljTWQINOVY
BdhsZs9pfhu/VVa3WDFadGGa8tM1dZe7jAnLGgFSdcQp9D4dfmGGvXguhWnTMnfL
sqqjVA+6ihTdO5lGd5Ek+zs0BjwP/4FvBndlmRiYIOm8Hqkw+8RR9gk0sUe4FDv8
mHC8G/7vlTg=
=dqQ+
-----END PGP SIGNATURE-----