🚀 go-pugleaf

mwm@contessa.palo-alto.ca.us (Mike Meyer) writes:

>I do agree that any language designed to be close to that class of
>hardware would look pretty much like C. After all, the features and
>constructs are dictated by the hardware, not the language designer.
>About all that's left is syntax. Sure, there are lots of places C
>could be improved, but they wouldn't lead to a radically better
>language, just one that was easier to read, without quite so many
>gotchas.

I was in a bookshop the other day, and was browsing through Tannenbaum's
book on Operating Systems. I was interested in his remarks on the design
of some early computers around the late 50s, early 60s. He said of the
Burroughs B5000 (ancestor of Unisys A Series desktop/mini/mainframe range),
that it was primarily designed by considering the software first. From
the beginning, it was designed to have all systems software written in Algol,
(which it still does, or Algol derivatives, such as NEWP, Pascal, etc).
(The B5000 also had one of the first commercial implementations of
virtual memory, 10 years before IBM invented it!).

His main point was that this was a successful attempt at developing a
machine by considering the software first, not the hardware, but as
soon as it was shown that this was possible, it was quickly forgotten.

Well, I won't forget it. It is the purpose of computer hardware to run
software. It is not the purpose of software to be crippled by the
compromises and limitations of current generation hardware.

So hardware design must be dictated by software considerations, not
vice versa, as is so apparent in languages like C.
-- 
Ian Joyner  ACUS (Australian Centre for Unisys Software)    ian@syacus.acus.oz
"Where is the man with all the great directions?...You can't imagine it,
 how hard it is to grow, Can you imagine the order of the universe?" ABWH
Disclaimer:Opinions and comments are personal.