#AmigaDOS Philosopy
2 messages in this thread
Bill,
If indeed arguments were passed on the stack in order of the documented
conventions, then there would be essentially no difference between documenting
in assembler, Pascal, or C. A function could be documented as:
returnvalue = function (arg1,arg20
and the assembler part of the docs could simply specify the order of the args,
and where the returned value could be found.
I was speaking with an Amdahl CE/SE today, who happens to own an Amiga, and
mentioned this discussion. He agreed with me 100%, and stated that it is
immediately obvious to him that anyone who wanted it documented in higher level
terms has never had experience with IBM 360/370 systems. Can you imagine
documenting a mainframe in PL/1?
Regards, Larry.
Yes, I can imagine documenting a main-frame in PL/I. It would beat the &^%$
out of most of IBMs manuals. The real point I'm trying to make is that we have
coddled the assembler person for TOO long. A problem is that most hardware
manufacturers are hardware manufacturers. They generally see software as a
neccessary evil to sell the "real stuff". So, they spend the absolute minimum
– 20% and push the stuff out the door. And then, of course, "we all know that
compilers generate inefficient code" syndrome from all the people who've been
around since the late 50's when nobody knew what a compiler was. So the
assembler gets written, then the OS (in assembler of course, 'cuz there is
nothing else), then the compilers (only the budget on them is generally
absolute minimum – 50%). And you see in front of you, computing in the USA
Today. (flame down)