Relay-Version: version B 2.10 5/3/83; site utzoo.UUCP
Path: utzoo!decvax!decwrl!ucbvax!NGP.UTEXAS.EDU!mknox
Newsgroups: mod.computers.68k
Subject: Re: DRI concurrent DOS for 68K
Message-ID: <8606230059.AA05340@ngp.UTEXAS.EDU>
Date: Sun, 22-Jun-86 20:59:00 EDT
Article-I.D.: ngp.8606230059.AA05340
Posted: Sun Jun 22 20:59:00 1986
Date-Received: Mon, 23-Jun-86 06:29:58 EDT
Sender: mwm@ucbvax.BERKELEY.EDU
Organization: The ARPA Internet
Lines: 48

   The CDOS-68K has been actually running for about a year and a half now.
It was done BEFORE the CDOS-86 for the PC.  [Actually, it was suspose to be
done in parallel, but it proved icredibly easier to write than the 80186
version.]  The idea was that the two systems would be compatible in design,
so that DRI could hit the market with an 'across the board' package which
vendors could write their products for and sell to both PC and 68000

    I have the full documentation on the system.  It is not a bad design
at all, although quite a departure from CP/M.  A utility shell was planned
so that CP/M-68K programs could also run under CDOS.  Basically, CDOS
did all the things that one would have done for a 1975 mainframe, if one
had the power in a micro.  It has print spoolers (REAL ones, multiple
programs driving a common spooler), multi-tasking, multi-user.   GEM
was designed to sit on top of CDOS.

    Hardware requirements were flexible.  An MMU was extrememly desirable,
but not required.  [No MMU means no protection from other users, and 
concerns with memory fragmentation.]  Without an MMU, programs were stored
in the new COMPRESSED format now used under GEMTOS on the ATARI.

    Now the down-side:  This product is a serious point of contention between
myself and DRI.  I worked with DRI on the early Beta of CP/M-68K, and (as
a strong OEM supporter of their 68000 effort) on the additions needed to 
make up a worthwhile system for the 68000 (CDOS-68K).  This was done
without compensation from DRI, because I fealt it was needed, they were
short of money, and it was important to myself as well that they get a GOOD
68000 product in the market.  

    Then they got even shorter of money.  While waiting for the beta of
CDOS-68K, I got very little support from DRI for CP/M-68K (although they
were always willing to cash my royalty checks I sent them each month).
Finally (two years late) they finished CDOS-68K.  I said, "where is my
BETA?"  They said, "We decided not to do that.  We will send you a copy
if you will send us $200,000."  What I said is not printable.

    I am STILL after them to change their policy.  The people I worked
with have been laid off (mostly), and the research facility here in Austin
which developed the system has been closed.  I *DO* believe CDOS-68K can
be a good 68000 system.  But it *MUST* be priced where it can compete
with the more established systems, and it *MUST* have support.  Motorola
paid several million to have DRI develop the system.  There is no 
development cost to recover.  DRI MUST show some sign of reason.  I imagine
CompuPro has managed to get a somewhat smaller price, but unless they can 
market the system cheaply enough, and attract language/application vendors
in quantity (something CP/M-68K has yet to achieve), they I can only
wish them luck.