🚀 go-pugleaf

In article <41d0sk$70n@meaddata.meaddata.com>,
Christopher C. Wood <christw@meaddata.com> wrote:
>In article <41astd$53u@steel.interlog.com>, daveacg@interlog.com (Dave Glue) writes:
>
>|> True.  Pre-emptive is more efficient and smoother, which is why it was
>|> added into 95.
>
>An interesting claim.  My experience is that the mouse on my
>workstation becomes unresponsive about as often as the mouse on my
>mac.  Preemptive multitasking systems can get CPU-bound and (with
>demand-paged VM) thrash... Preemptive multitasking adds overhead.
>
>It usually makes the job easier for the programmer (protecting
>critical sections is harder).

Yes, but as I have stated before, if you take advantage of the
cooperative model, and don't protect critical regions with explicit
constructs such as monitors, semaphores or message-passing, your code
will have to be fixed when you move it to a system that does
pre-emption, and possibly real concurrency (SMP, or networking).
Semaphore-based concurrency control will work under pre-emption and SMP
(with some modification to the semaphore code, perhaps).
Message-passing synchronization can work across machine boundaries in a
distributed system. 

Why make the job easier for lame programmers?  Give the job to someone
who knows what he or she is doing.


Gee, why don't we all go back to coding in 6502 assembly? Those Apple II
games sure rocked!
-- 
Finger c2a192@keats.ugrad.cs.ubc.ca for PGP public key. Or contact
one of the public key servers -- the key id is 0xD3C7995D.