Richard Erlacher wrote:
> The short verson is that you're right! I haven't encountered even one
> problem with the libraries I used under CP/M, and that is due, in part, to
> the fact I seldom used them.
If you are using C, you must use libc . If you are using FORTRAN,
you must use libF77. And finally, you need to use the optimizer/code
generator of the corresponding compiler. I had MS Fortran powerstation
"optimize" out operations by totally failing to execute them! (well,
actually, disassembling showed that it actually forgot to save the
result of the operation...) if you grouped parentheses in a certain way.
Then the thing leaked memory... even when using static allocation!
> Well, the OS was written to be used with applications, including, as you
> say, certain compilers and other tools. If they don't work the way they
> should, there should be legal consequences. I take great pains to test my
> hardware and verify the documents I deliver with it truly represents what
> happens in normal operation. If Microsoft, or any other vendor for that
> matter, fails to do that, it's because too many people let them get away
> with it. If there were lots of complaints, phone and written, they'd
> gradually close those gaps. People are too lazy to complain, or they know
> their complaints are unwarranted.
Unless, of course, the only available product in the market for doing
what you want is made by microsoft. Fortran powerstation failed
spectacularly as soon as DEC came up with Digital Visual Fortran
(a nice product, despite the name). And then there was the issue of
how MS treated former fortran powerstation users... failing to provide
fixes or refunds for a flawed product when they retired it.
--
Carlos Murillo-Sanchez
428 Phillips Hall, Electrical Engineering Department
Cornell University, Ithaca, NY 14853
Received on Fri Apr 07 2000 - 12:35:16 BST