28 Apr
2002
28 Apr
'02
3:23 a.m.
Yes, the enhancements that finally found their way into to the Model 4 were
available from third parties (with the possible exception of the display
format and speedup) for the Model I. We offered our own versions in a design
alternative done under contract to Tandy as a Model 3 design, in the very late
'70's, but it cost them $5 per unit more than the lowest-cost alternative, so
they rejected it. They paid for the work but didn't use it.
It's too bad it took them so long to realize that having two operating systems
available rather than just one would help their sales rather than hurting
them, and that a conventional display made a lot of sense, and that having the
slowest computer on the market might be hurting their competitive position.
By the time the M4 came out, most of the Z80 systems people wanted were
running in the 6-8 MHz range and cost considerably less than the M4. Tandy
management was just not able to keep pace.
Dick
----- Original Message -----
From: "Tony Duell" <ard(a)p850ug1.demon.co.uk>
To: <classiccmp(a)classiccmp.org>
Sent: Saturday, April 27, 2002 3:18 PM
Subject: Re: expansion differences (was Re: Micro$oft Biz'droid Lusers)
expanded
from
even
Which, depending on the task could be done using a row of LEDs, a 7
segment display, a printing terminal, or....
> > Well, if you're just starting out you have to get a lot right just to
run
> said was that the basic unit wasn't much more
than a toy, since the EI and
> disk drives didn't come out until some time after the TRS-80 became
avaiable.
> The Radio Shack experience was pretty much the
first one of its kind, i.e.
> they put out what they thought met their "most likely to sell" criteria,
> though they were pretty wrong about the concept in several respects, all
> associated with packaging. They figure that out and fixed most of the
> problems in the Model 3. They didn't fix all the problems, of course,
such as
that inane
display format and the slow cpu, but, they got the packaging
somewhat better.
The Model 4 did fix the display format (80*24), the CPU speed (IIRC it
could run the CPU at 4MHz) and it even let you switch out the ROMs to
have a 'RAM-only' memory map to run CP/M. Of course it was rather too late.
Actaully I prefered display dimensions being nice round numbers like 16
and 64. I could never really understand the reasoning behind 80*24
(except that it was a 'standard').
> > I really can't see why you claim the Apple ][ was designed as a computer
> > and the TRS-80 (say) was not.
> >
> The obvious distinction is in that the Apple ][ was intended to be with cards you
could buy. Like the other "toy" vendors, which came along
Err, and the CoCo (which is the TRS-80 which started this discussion --
please don't arbitrarly move to a different machine!) was intended to be
expanded by cartridges you could buy (like the disk controller, RS232,
speech/sound, etc).
> somewhat later, RS figured you should buy expansion hardware exclusively > them, so they didn't provide a generalized
expansion interface, but,
rather,
Eh? the CoCo expansion slot was the 6809 bus. Like the Apple slots were
the 6502 bus. Both machines had a slot pinout, etc, that was designed for
that machine, and both manufacturers would probably have liked it if
you'd bought all your add-ons from them. But none-the-less both
manufacturers did tell you the pinout and signal specs so you (or any 3rd
party company) could make add-ons if you wanted to.
I really can't see the difference between the CoCo and Apple ][ in this
respect.
> > Apart from pre-packaged business systems, very few micros came with word
> > processing software. It was always an optional extra.
> >
> Yes, but if, by way of comparison, you bought an Ampro "Little Board"
which
> only preceeded the CoCo by a few months (?), you
had your choice of
several,
Sure, but none of them were free (AFAIK). Wordstar was not cheap back
then...
And the CoCo did have software available essentially when it came out
(or at least _very_ shortly afterwards). And it wasn;t that expensive
either.
> > So, you needed to buy the Little Board, a serial terminal, at least one
> > disk drive (I can't believe $200 included the drive), and probably a PSU
> > and a case. And then get it all working. Fine if you knew what you were
> > doing. Not so good if you were starting out.
> >
> > And presumably using Modem7 to transfer programs from 8" disks (the
> > _only_ standard CP/M format) to the Little Board implied you had another
> > CP/M computer with 8" drives. You know, this is starting to sound
expensive.
> >
> Well, if you bought a new terminal, it meant an outlay by 1983 of
somewhere
on
the order of
$550. If you got one second-hand, $100 might do it. If you
Exactly. Rather more than a CoCo.
In other words, if you had the money there were better machines than the
CoCo. Nobody disputes that. But I would guess that getting a Little Board
system up and running using entirely new parts (Little Board, Terminal,
Drives, Case, PSU, etc) would cost around $1000. The CoCo was under half
that. Yes, by the time you'd added drives, monitor (rather than using
your existing TV), etc to the CoCo you'd probably spent as much. But you
could spend the money a bit at a time and have something that was useful
along the way. And of course in the end you did have a multi-tasking
computer with a much nicer processor than the Z80.
> >
> > Sorry, I have to disagree with you there. Many of my computers don't
> > have any form of video output. The think I
look for when choosing a
> > computer is the _processor_ and _OS_.
> >
> I've long (very long) considered a display of fewer than 24 lines of 80
> characters pretty useless. True, 80 column lines in pages of 24 date back
to
As I said, having used many systems with less video capability than that
for real work, I can't agree with you.
> > Seems reasonable. But that's because the purpose of a Hi-Fi system is to
> > produce sound with minimal distortion, and most of the distortion comes
> > from the speakers. The purpose of a computer is not just to output
video.
It's to do processing as well.
True, but it's got to tell what the outcome of your processing task was.
> > that monitor. All the bus connections
correctly wired and no shorts. CPU
> > clock running. Memory map as you (and the monitor ROM) expect it. ROM
> > correctly programmed. I/O devices where the monitor expects them. And so
on.
> >
> > None of this is impossible, or even that difficult. But for a first time
> > project there's a lot that can go wrong, and you might well then not
know
> > how to look for the problem. If you started
with a CoCo (or whatever)
you
> > had a base machine that would work. If your
device stopped the machine
> > from even starting up it was a fair bet you'd got a short between 2 bus
> > lines. If it didn't repsond to the right addresses, you'd mis-designed
> > the address decoder. And so on. It was a lot easier to get your first
> > hardware project working that way.
> >
> removing the CPU and using switches to set the addresses was an easy
enough
> way to check whether a CPU communicated properly
with an I/O decoder and
the
Oh, sure... I've done that many times. That's not the point...
When you know what you are doing, there are plenty of ways of
fault-tracing using minimal equipment. But when you're starting out you
probably don't have the ability to work out things like that. Having
something to start from that you know works is a big step forward. Or at
least it was to me.
This was one thing that was so nice about the BBC micro. It had an 8 bit
user port (half a 6522) and a 4 channel ADC built in. Now, we both know
how easy it is to link up a 6522 or an ADC chip to the 6502 bus. But I
can assue you that the fact that these interfaces were built-in on the
Beeb enocouraged a lot of users to link up relays and simple sensors,
etc. And to actually try some interface. They probably wouldn't have
bothered if they'd had to wire up the interface chip themselves first.
And they certainly wouldn't have tried making a computer from scratch.
> I guess I consider it pretty simple because that's how I got my first
computer
> working. I took a memory map from an existing
architecture and, because I
was
Actually, I debugged my first homebrew like that as well. And used an LED
+ resistor to monitor bus signals. From the brightness of the LED I could
estimate the mark-space ratio of the signal (and certainly spot lines
that were always high or low).
I am not saying it can't be done. Just that few beginners are going to do
it.
Another data point. Look at all the projects on the internet that connect
to a PC printer port. Now, making an ISA card is almost trivial, and in
some cases it's actually more work to kludge the interface onto the
printer port than to make a special ISA card for it. But none the less,
people prefer to just plug into an external connector on an existing
computer.
I was hoping to avoid that ...
I see, you (as usual) can't be bothered with facts....
-tony
resident ROM was handy, provided you could make it go
away.
You could. Look at the SAM data sheet.