Date : Thu, 10 Nov 1983106:56:00-PST
From : lauren@rand-unix
Subject: Erroneous information from hp-pcd!craig about MARC
I sincerely hope that this will be my last message on this topic.
I don't know what hp-pcd!craig has been smoking, but his information
regarding MARC is absolutely and totally wrong and confused.
There isn't any "George" at Vortex. I AM VORTEX. VORTEX IS ME. Period.
I will NOT be selling or distributing MARC in any manner.
The MARC software project has been terminated.
MARC was designed only for the 8080/Z80 processors and there
have never been any plans to distribute a MARC for the 68000 or any other
processors. In point of fact, the overwhelming percentage of software
in the MARC software package is written in a non-standard 8080 assembler
and is most decidedly NOT portable in any manner.
To be blunt, the system was not really usable as other than a toy.
Performance with floppies was miserable and could not be reasonably improved.
Even with hard disks, many operations were extremely slow. The system
could NOT make use of additional memory over 64K in any manner, and
the useful workspace for user programs ended up being only around 30K,
sometimes even less. CP/M compatibility did not function properly
for about 75% of currently tested CP/M programs.
The MARC software package is fundamentally limited by its original
design parameters, and has no future beyond hardware which
is rapidly heading into oblivion -- and, as I stated, it doesn't
work well enough even on that hardware.
There are a variety of software products from various vendors on
the market which can provide much of the MARC functionality
in a much more reasonable manner, and which won't ignore the entire
base of existing CP/M software in the process. Microshell and Software
Tools are two obvious examples of reasonable approaches to the
problem of providing such an environment on limited machines.
There are also packages which can make effective use of bank-switched
memory and provide for much faster disk access, which should help
to provide functionality for that hardware which MARC could not and cannot
provide.
MARC was a good effort but is just too fundamentally limited by the
underlying hardware base for which it was designed and written.
It is just "too much" for such hardware -- the operating system
takes up so much of the memory and disks that there just isn't
anything reasonable left for the humans! Also very important
is the fact that MARC's being written mostly in 8080 assembler
made it difficult to maintain and modify and essentially impossible
to take forward into the future in the rapidly changing micro marketplace.
You might be interested to know that of the people I've talked to about
the termination of the project, the vast majority admitted that they
were planning to try upgrade to newer hardware (usually with lots
more memory and usually running a fullblown multiprocess Unix
or real multiprocess Unix look-alike system) in the near future.
Most of the people (few as they were) who sounded the most disappointed
were those with hardware that would not reasonably run MARC in any case.
However, the bottom line is that bugs and poor performance would
require so much more code to fix properly that the remaining memory
space would be made even smaller and less useful!
I don't sell *or* distribute software with which I am not happy.
I never sold a single copy of the MARC software package because
I refused to send out buggy and limited software. It doesn't
matter whether the package was going to cost $0 or $500, I simply
refuse to distribute software with which I am dissatisfied.
I've spent a large amount of time on the project, and
I'm not happy about the final outcome -- but it's time to
face reality on this topic. It was fun trying, anyway, but I've
made my decision and it is final -- I need to get on with my life
and try to make a living!
I really have nothing more to say about this. That's all, folks.
--Lauren--