[Coco] The Tri-Annual CoCo 4 Thread

[Michael Robinson]

On Wed, 2014-02-12 at 21:58 -0500, Brian Blake wrote:
> Hear, hear!
> 
> Well said Bill.
> 
> ------------------------------
> www.tandycoco.com
> www.tandycoco.com/forum
> On Feb 12, 2014 9:51 PM, "Bill Loguidice" <bill at armchairarcade.com> wrote:
> 
> > Is it possible you're misremembering how much your CoCo 3 gift was? Even
> > when it was brand new, it would have been incredibly difficult to come up
> > with a $2000 configuration, no matter how much you loaded it up. It really
> > wasn't designed to exceed $1000 even in its most extravagant configuration.
> > Also, no one is expecting the CoCo 4 to have any type of mainstream
> > success. It won't. It won't even have Raspberry Pi success, which is the
> > biggest hit for these types of hobbyist systems (and a lot of that has to
> > do with the extraordinary price). The "new CoCo" (I hesitate to call it a
> > CoCo 4) has to be targeted to the CoCo fans like the ones on this list,
> > and, preferably (more maximum impact) the Dragon fans. I'd say sales of a
> > few hundred units at $150 - $300 would be reasonable if it hit enough
> > checkpoints on most wishlists, but numbers beyond that would be wildly (and
> > baselessly) optimistic. There's a LOT of competition out there in this
> > hobbyist category, and has been stated time and again, not a great deal of
> > nostalgia for the CoCo to help drive anything remotely like mass
> > production. Obviously we're all trying to change that (the book, the new
> > homebrew projects, etc.), but we've got a LONG way to go.
> >
> > -Bill

My COCO 3 was one of the last ones made and it was at least $1k,
probably $2k.  With the attitude no interest can be drummed up, of
course there isn't a COCO 4 yet.  You've gotta dream big enough to 
get anywhere, yet not so big that nothing can be done.  I looked
for a multi pak as I think I gave mine away.  There is only one on
E-Bay and they want over $200 for it.  Yikes!  If it sells for that,
that'll put the lie to there being no interest in the COCO.

Radio Shack's approach to selling the COCO was terrible, and it still
succeeded for over a decade.  A little bit of decent marketing, good
engineering, and fair pricing should make a COCO 4 sell easily.

Here is why a COCO 4 is needed and desirable.  Color computers were easy
to program compared to the alternatives back in the day and a COCO 4
could likewise be a dream to program.  Color computers were personal
systems and very usable without flaky media.  There is a COCO 4 basic
already, but no machine to run it on native.  Emulating on today's
less than desirable "modern" PCs just won't cut it.

Instead of a ROM, Nitros9 on a 32+ gig SSD might work.  An enhanced
version of Basic and a simple command, say nitros9 to get to a gui makes
sense.  A 128 gig SSD can be had for less than $100.  Surely a smaller
one can be had for less than $50.  The processor can be designed using a
xylinx perhaps.  You need a GIME chip and 6809E equivalent to achieve
COCO 3 compatibility.  Should be able to add 16 bit and 32 bit
instructions allowing one to access up to 4 gigs of ram.  Modern HD
televisions and modern LED flat panels can do 16 million plus colors.
Instead of disks, switch to creating special files on the SSD and
serving them as if they were local disks.  Another option is to
implement USB so the COCO 4 can read USB memory sticks.

A COCO 4 is interesting if it is a pleasure to program, downward
compatible, and far more powerful than it's predecessors.

Right now, games like Gauntlet II aren't playable for most people.  A
dwindling number of people have good COCO disk drives where Gauntlet
II's copy protection prevented making backups.  I have dsk images of
Gauntlet II, but the best I can do is try to use an emulator on an
IBM compatible.  It is miserable to play Gauntlet II under emulation.

Don't make a COCO 4 that plugs into the television, make it a laptop.
Compete with Google's Chromebook.

If I'm not mistaken, the Color Computer 3 could connect via ethernet to
a local area network.  No reason a clone can't do so.

Here is what I see as reasonable...

Build an advanced 6809 processor/GIME chip replacement that runs at
minimally 800 Mhz.  Add to this processor new instructions like the old
ones that allow adding, multiplying, loading, storing, and subtracting
on 32 bit registers.  Add new pokes that allow the speed to be adjusted
all the way down to what the CoCo 2 ran at, to high speed on the CoCo 3
to 800 Mhz+.  8 bit pic micro controllers already run at 8 mhz.  Don't
know about the 32 bit series.  Allow the memory bus to be sped up so 
you can access up to a gigabyte of memory in 100 cycles or less.  In
short, build the 6809 using modern technology and add 32 bit
instructions speeding the bus speed up a lot.  If at all possible,
achieve a Gigaherz or better without active cooling.  Don't build a 
CISC processor to the extent possible.  A RISC processor will probably
be faster.  Taking the Hitachi 6309 or the Motorola 6809E, what should
the 32 bit grandchild look like?  A program written for grand dad should
run on the grand child, I know that much at least.

My first program was written on a color computer.  I hate that the COCO
has been abandoned.  No other computer since has been as much of a
pleasure to work with.