Date   : Wed, 03 Jun 1987 02:42:29 GMT
From   : phoenix!pguhatha@princeton.edu (Puragra Guhathakurta)
Subject: zcpr3.3

       ZCPR3.3 : some first comments

                               Peter Teuben,
                               June 2, 1987

This concerns ZCPR3.3, the latest update in the line of 
command processors for CP/M, and my first experiences.

We have all watched ZCPR grow over the past years to a fairly
grown up module of what we now call the Z-system. None of
the original modules from Digital Research are still used,
the CCP being replaced by ZCPR and BDOS replaced by ZRDOS
(often we write our own BIOS's).

ZRDOS can be only be obtained commercially (Echelon), but 
ZCPR(3.3) can in principle be obtained for free (although at
Echelon you have to pay $4x, but I guess they will accompany
my order with a nice installation manual).

However, I was impatient, and got the 100k of initial startup
sources from SIMTEL20. (That's for free, as most BBS's are, I
presume)

For me this kind of installation is just an evening of work,
and very amusing to do so. I installed ZCPR3 last summer in just
an evening, including all modifications to the BIOS (thanks to the 
excellent documentation by Richard Conn).

Last september I witnessed a rough discussion between Richard Conn
and Jay Sage, on their use of the appropriate assembler for the
'Z-Libraries', i.e. ZAS (Echelon) versus Z80ASM (SLR).

I also listened to Richard Conn's story on his building of ZCPR33, 
on last years Trenton Computer Festival (april '86), and now it is 
there!

To my great surprise the author of ZCPR3.3 turened out to be .... 
Jay Sage!

What ever happened to Richard? (are you listening Richard?) Or his code.

But, and this is now perhaps not a big suprise, I completely
failed in installing ZCPR3.3 (I thought it was advertised as having
an even easier installment as with V3.0, which I thought was already
pretty simple), because all of my 'reliable' compilers failed:
ASM, MAC, RMAC, M80, Z80MR. (ASM, MAC, RMAC couldn't help it, the 
new code is now written in Z80 opcode, which I can perhaps
agree with; what will happen to the poor old 8080 users, or did
they all die?)

Apart from a few trivial problems (who EVER came up with the idea
of replacing brackets (yes: ()'s) with square bracket's []????),
the author used a compiler where one needs now at least an 8 character-
length variable name detector. We all now that M80 exports only
6! I phoned SLR, the experts there told me that M80 V3.44 exports
6 yes (because of the .REL format), but can internally use 16
character lenght!!
I tried mine (V3.4 from 1980!!), and even a little test, but no way. 
Does anyone know of that (newer) 3.44 version, and is it true what SLR
told me. Other than that, M80 would probably have done the job (apart 
from two extra needed .PRINTX's and having MACLIB filenames in full and
uppercase). Even compilers with 7 character length variable names
will fail, as a quick glance over the code learned.

The end of the story you can probably guess: I dialed into Jay's
machine last nigh and found an order option for the SLR compiler (and Linker,
but one could live without this one). After reading the description
of the compiler, I'm really impressed. Like 5-10 faster than M80,
and so many more possibilities. It will be quite an improvement for
assembler programming, but I keep thinking of the funny stories I
read in these rebuttals last september.

I am looking forward to receiving my SLR compiler from Jay,
(Btw Jay, I have a Kaypro 4, DS DD)  and
will perhaps let you netreaders know what happened. After all, ZCPR 
is really a fantastic piece of software, and moreover, allows YOU to 
write and modify part of the operating system (&utilities), which is
not possible on many machines for a price like we pay for this

Let me know what your experiences are, do all of you know have these
fancy new compilers, like ZAS and Z80ASM? I am just old fashioned?
Afer all, at some stage we have to make use of the Z80 itself. Of
course if we extrapolate this, soon the Z80 users will find themselves
deserted, because ZCPR4.0 will be written for Z280 or HD64810 only.

Peter Teuben
Institute for Advanced Study
School of Natural Sciences
Princeton, NJ 08540
TEUBEN@IASSNS.BITNET or pjt@astrovax.uucp

(PS: I have no regular means of up/down loading to BBS's, in particular
the Z-nodes. I would appreciate if someone could distribute this)
----