[net.micro] Digital Research C compiler

avsdS:nelson (01/20/83)

Last night the Homebrew Computer Club (Silicon Valley) was treated to a
very interesting presentation by Tom Rolander, a V.P. of Digital Research,
and author of MP/M. In spite of DR having used PL/M for much of their work,
having written PL/1 for 8080 and 8086, having acquired Pascal/MT (bought
the company), they decided to use C for all future op/sys development!

DR actually tried to write a portable version of CP/M using Pascal -
"gross" sums up Rolander's opinion. For other reasons they had given
up on PL/1 by this time. Then they tried every C compiler available
for 8086 and found "serious" problems with all. So they wrote their
own, developing it on their VAX running UNIX (which Rolander says they
all like!), then wrote CP/M in C, and ported it to a 68000! He reports
that it takes less than 2 weeks to port it to a new processor! This
brought murmurs of appreciation from the audience (which is more hardware
oriented), who probably think this is testimony to the power of C.
It is, of course, but I suspect that at least on wizard is at work down
there, and that CP/M itself is no big deal to port.

Now for the really good news: CP/M-68k will shortly be released for $350
with the C compiler! I'm no big fan of CP/M (though I do use it), but it
does have its place. I applaud DR for their decision re C.

By the way, Rolander (in spite of being a V.P.) seems like my kind of
programmer - commments such as: "when properly abused C can be used as
a structured assembler", "I like writing obscure, exotic code" (with
a big grin on his face) tell you where he got his training.

		Glenn Nelson, Ampex, Redwood City