* Re: Military software computer language
[not found] ` <33E65CE2.4908@unibw-hamburg.de>
@ 1997-08-04 0:00 ` Mukesh Prasad
0 siblings, 0 replies; only message in thread
From: Mukesh Prasad @ 1997-08-04 0:00 UTC (permalink / raw)
Roland Jakobs wrote:
[snip]
> When you ask these questions, you will understand, why there was a strong
> need for standardisation not only for a language definition itself, but
> for the compiler and the environment too, as in ADA.
Actually, I think one of the problems with Ada(-83) was its
attempt to define a common environment.
Even today, nobody knows what is a good common environment.
Nobody can say for sure what is the best way to do threads, IPC,
semaphores, shared memory... At best you might get "I like this
approach better".
When these problems all have obvious known solutions,
it might make sense to put this knowledge into a language.
Otherwise, it doesn't make sense to have the language
designer mandate one particular solution as the one that
everybody has to use in all situations.
Ada forced compiler writers to come up with schemes
for doing multitasking and synchronization. These
schemes went to a lot of trouble to meet with
the language's requirements, even if it meant
doing something completely alien to the underlying
operating system and coming up with something
much worse than the native solutions for that OS.
It is sort of like if Java had come out in 1985,
and tried to force all Windows developers to use and
be restricted to the widgets and programming
style of XToolkit.
^ permalink raw reply [flat|nested] only message in thread