[fa.info-cpm] Marc

C70:info-cpm (06/04/82)

>From Jeffrey@Office-2 Thu Jun  3 23:27:10 1982
This query seems to show up regularly; its time for another round.

Would the appropriate parties/party please give us a little status report
on MARC?

Over the past couple of years there has been a good deal of enthusiasm
for this system, but little real information.  Will it ever come
to be?

Thanks,

Jeffrey Stone
Menlo Park, CA
-------

info-cpm (11/30/82)

>From npois!harpo!decvax!goutal  Tue Nov 30 01:11:31 1982
Received: by UCBVAX.BERKELEY.ARPA (3.227 [10/22/82])
	id A12767; 30-Nov-82 01:22:37-PST (Tue)
To: harpo!npois!ucbvax!C70:, info-cpm@Ucb-C70
Via:  Mit-Ai; 30 Nov 82 2:14-EST
Via:  Brl; 30 Nov 82 2:16-EST
Via:  Brl-Bmd; 30 Nov 82 2:42-EST

What IS it?

info-cpm (12/03/82)

>From vortex!lauren@Lbl-Unix  Thu Dec  2 21:22:17 1982
To: clark.wbst@Parc-Maxc
Cc: INFO-CPM@BRL
Via:  Lbl-Unix; 2 Dec 82 20:08-EST
Via:  Brl; 2 Dec 82 20:15-EST
Via:  Brl-Bmd; 2 Dec 82 20:20-EST

Greetings.  Yeah, your description is pretty accurate.  I like to
consider MARC to be sort of a V6/V7/4.1bsd hybrid -- over time I've
tried to add those features from V7 and Berkeley that I particularly
liked (and that made sense on a very small computer!)

Actually, the filesystem is very closely patterned after V6, primarily
because the V7 layout would probably have taken up too much overhead
space on people's small disks.  The MARC block size in 256 bytes, again
in keeping with the smaller scale of the system.  Many of the "obvious"
Unix utility commands are around -- in some cases (e.g. "cp", "mv") these
are patterned after V7 rather than V6.  The system calls include
"tell" (which of course is "late" V6).  Login startup files, private
bin directories, and similar V6/V7 features are supported.  We now
also have a "make" command (very similiar to V7 Unix "make" but not
as complex).  Very useful, however.  I have special versions of "sq", and
"usq" for MARC, but have yet to receive final permission to distribute
them with the basic system.

The C compiler is indeed a special version of BDS C, with long and
float subroutine packages.  The library support for MARC from the compiler
is extremely extensive.  The MARC stdio library is virtually identical
to the K&R Unix standard (makes life *alot* easier!).  The cp/m emulator
runs, and tries as hard as it can, but given the variety of insane
cp/m programs (some of which open 20 or 30 files w/o bothering to close
other files, etc.) it will never be able to handle EVERY program.  However,
it seems to do fine with most editors (like Wordmaster) and most of 
the language translators I've tested (MAC, M80, L80, TLC LISP, etc.).
A MARC version of MINCE is also available -- I couldn't get along without it.

Indeed, 'tis true that "vortex" is a Z80 running MARC.  In fact, I'm
typing this message directly into the MINCE that interfaces with my mail
generation software.  I *may* include my basic mail handling software
in the MARC package, or it may be included in a utility package later on.
I have *not* made any decisions about how I will handle release of my
MARC/BDS C UUCP code, except that I know it will not be part of the basic
MARC package.  (Whaddaya want for $250?)

MARC is not available yet.  It has taken a LONG time to get this far, and
hopefully there won't be too much more of a wait.

--Lauren--

info-cpm (12/04/82)

>From ELIOT@Mit-Mc  Fri Dec  3 23:09:30 1982
To: vortex!lauren@Lbl-Unix
Cc: info-cpm@BRL
Via:  Mit-Mc; 3 Dec 82 18:30-EST
Via:  Brl; 3 Dec 82 18:39-EST
Via:  Brl-Bmd; 3 Dec 82 18:44-EST

The one question that comes to mind is: with all these features, and what not,
(is it written in C or ASM, if asm 8080 or Z-80?) how big is it?  It must
be huge!

						-Eliot at Mit-DM

info-cpm (12/06/82)

>From vortex!lauren@Lbl-Unix  Sun Dec  5 23:47:19 1982
To: RMS.G.BANDY.MIT-OZ@Mit-Ml
Cc: INFO-CPM@BRL
Via:  Lbl-Unix; 5 Dec 82 22:56-EST
Via:  Brl; 5 Dec 82 23:17-EST
Via:  Brl-Bmd; 5 Dec 82 23:26-EST

I'm glad you've seen MARC, but please don't spread misinformation
around!  MARC does NOT require "BIG in-memory buffers".  All existing
MARC implementations run within 64K bytes.  I have run MARC successfully
in as little as 48K, but admittedly that was a bit cramped for getting
anything useful done.  You DON'T need a "big winnie" either.  Hard
disks are nice on almost any system, but MARC will run fine even on
single-density 8 inch floppies.  Since there are so many utilities
(resulting in a pretty large /bin) it's nice to have double-density
floppies so that you can fit all the utilities onto your root disk, but
even running single-density it is fairly trivial to simply put most
of your "infrequently used" programs on a different filesystem (floppy).

I've never been able to figure out exactly how to benchmark MARC, but
on my single-density system, the access times generally seem to be
at least as good as I get for CP/M, overall.  Certain tasks might
take a bit longer and some a bit shorter, but overall the feeling
is that they are very similar.  Note that MARC does NOT handle disk
accessing (internally) exactly the same way as UNIX, and there have
been conscious design decisions made with an eye toward good performance
on smaller disks (for example, file block buffering is handled differently,
and the shell search path is essentially reversed from that of standard
Unix).

Most of the test sites ran double-density systems with 58-64K bytes,
and they generally reported that performance was very good, even 
with their simple floppy systems.  Obviously the quality of your
BIOS, and similiar factors, will impact overall performance, and a
hierarchical filesystem implies more time spent performing certain
tasks (and less time performing others), but overall the reaction
has been very good.

--Lauren--

info-cpm (12/09/82)

>From hplabs!hp-pcd!jay@Ucb-C70  Wed Dec  8 22:56:30 1982
Received: by UCBVAX.BERKELEY.ARPA (3.227 [10/22/82])
	id A29245; 8-Dec-82 23:05:41-PST (Wed)
To: hplabs!ucbvax!info-cpm@Ucb-C70
Via:  Mit-Ai; 8 Dec 82 18:33-EST
Via:  Brl; 8 Dec 82 18:48-EST
Via:  Brl-Bmd; 8 Dec 82 18:57-EST


Hello,

	My name is Jay Phillips.  I work for Hewlett Packard's Personal
Computer Division in Corvallis, Oregon.  I've been a CP/M hacker for 
several years and a UNIX hacker more recently.  I was also responsible
for the software portion of HP's CP/M system for the HP86 and 87 computers.
I have been following with great interest the news traffic about MARC.
I was wondering if there is any way for me to get a copy of it, possibly
as a test site.  I would like to bring MARC up on our CP/M system (it is
a Z80 card which plugs into the HP86 or 87 mainframe, similar to the apple
Z80 softcard.  A large majority of our customers are technical/engineer
types and I feel that there would be a good demand for MARC running on our
machines.
	Please send me a note letting me know what you think about this.
My UUCP address is below.
	Thanks much.

		Jay Phillips

		ucbvax!hplabs!hp-pcd!jay

P.S.	If you are familiar with Berkeley UNIX you no doubt are aware of
	a massive game called Rogue.  We were intrigued by it here and 
	have been working on a version written in BDS C for CP/M.  It's
	roughly 90% done.  Unfortunately, I can't convince my boss to 
	fund it as a project so it only gets worked on in my spare time.
	I would also like to bring it up under MARC.

					Jay

P.P.S.
	If U.S. Snail or telephone is more convenient, the particulars
	follow.

	Jay Phillips
	Hewlett Packard
	Personal Computer Division
	Lab 3
	1010 N.E. Circle Blvd.
	Corvallis, OR  97330

	(503) 757-2000 x4124

info-cpm (12/10/82)

>From decvax!harpo!seismo!presby!aron@Ucb-C70  Fri Dec 10 14:50:51 1982
Received: by UCBVAX.BERKELEY.ARPA (3.227 [10/22/82])
	id A16181; 10-Dec-82 14:56:33-PST (Fri)
To: seismo!harpo!decvax!ucbvax!info-cpm@Ucb-C70
Via:  Mit-Ai; 10 Dec 82 6:57-EST
Via:  Brl; 10 Dec 82 7:18-EST
Via:  Brl-Bmd; 10 Dec 82 7:41-EST

Marc sounds wonderful to me.  Maybe we should leave lauren alone so he
can bring it to market quicker!

					aron shtull-trauring
					seismo!harpo!presby!aron