2022-08-25 14:17:01 UTC
also think how our past world could have been designed.
I really regret some things, such as the failure of Novix. That the competing UK forth
chip was going be a twin chipset, which has a cost affect.
Looking back, the Novix/RTX design was excellent for a Commodore 65. On these
base machines, you didn't really need all the higher end OS features of the 68000.
Particularly machines like the Sinclair machines. It would be a great match, particularly
if you integrated some graphics features ( like multiple colour depth
resolution character tiling, scrolling, and some shared memory sprites and a basic
integrated chipset (or at a home computer level). You could loose 3 out of 4 cycles
to shared memory access through a memory hub function, and still be
competitive. The GEOS GUI package, could have made an excellent home computer. A
misc processor, even better, even on the original Atari VCS. The Sega Mega drive or
Genesis, could have flown, by that time, Computer Cowboys could have offered them a 32
bit shboom processor, a leading design. With all I know now, the time was ripe
for such a thing to happen.
Back in the day, many firms came around knocking on doors of companies, peddling
advanced gaming technology to firms. Some of the leading chipsets in game consoles,
had been peddled to multiple companies. The companies who adopted them, often
survived or led the field. But, back in 1983, there were only the Commodore 64, Nintendo,
Atari 7800 chips etc that led the way at the low end. Commodore 64's chip had
some draw backs, and something not too much more complex was needed
for the Commodore 128.
The Atari 7800 was a paid for design by a company that designed unofficial mods for
Atari Arcade games, they used arcade like technology, and was well and truely out
there, but hobbled by poorer system resources, and processor. A Novix, there,
wow. You have yourself something to come under the Atari ST and out compete
Commodore had bought the Fab firm, MOS Technologies at a bargain price, and used it
to undercut the competition. But, to stay up to date, you need to upgrade your fab, lots of money, so you can provide
better chips. They had problems trying to design a 16 bit 6502 replacement. A Novix, or Misc, would have
been something, which could get better performance without updating the fab. I
forget the name of the engineer there, who came up with the Commodore 128
design (which was supposed to use a much better graphics chip, but he was talked into
using an in-house part, which turned out to be much worse. I think the chip he wanted, might have been the Motorola
one people think used Amiga technology through some collaboration, the specs are
similar). He was just short of inventing the first integrated home computer shared memory chipset on the plus 4
range (which type of chipset was probably by Acorn first, who did the Electron and ARM). The issue is, that they
could have released earlier with a Novix. The 64 was really a stepping stone to
something more like latter 16 bit home computers.
One issue that really stuffed everything, is the game console crash (and at one
time home computers). This stopped the release of of the Atari 7800, and a number of
advanced consoles, and ET was hidden in the desert (joke about the ET game cartridges
:) ). But, Nintendo made heaps of money in Japan, releasing a better machine,
that was worth buying. Companies had in their hands a formular to get consumers
buying again, and pull the companies through, but many didn't release the better
machines, and suffered. So despite the bleakness, it was still an excellent time for a
Novix like cpu to be picked up by the knowledgeable (I'm not saying here the Novix as was,
but a version with bugs fixed, and some more new features. 40xx+/60xx+?).
The interesting thing is, maybe you could have a Novix derived design in a pin
compatible socket on the 64 design, or extended by 8 bits of the data bus. It maybe
have been possible to use, or largely use, the entire Commodore 64 motherboard
and other components, which Jack would have been very happy about. However,
replacing most chips and integrating functions into one socket free's up space to put
memory. Even today, FPGA's are put into modules, which go into these sockets, as
replacement repair parts.
Now, the Atari VCS with such a chip, but better misc, from the beginning, would have
allowed much better performance and clearer upgrade path. It's graphics modes are software driven, so
a better CPU would help. I've come up with a insane theoretical design for an
game engine for the VCS, which could enable you to develop games of the level of
Sonic the Hedgehog on it, which would look like the version on the Sega Master
System. With a misc or Novix like design, things could be even better. There is
absolutely no real market for something like that on the VCS today, but I am thinking
of including it on the retro gaming chip I have planned, and a new object orientated
programming system to complement Forth. The retro chip was to use massive
compression techniques but, every redirection to the eventual data destination,
takes up a cycle or so, or requires parallelism (tiling and tile reuse, and reusing sprites
images, are all basic compression techniques used
back then, so it's not totally unreasonable). I designed my theoretical GA144 renderer
around this problem, so, on average, nearly 0 extra cycles to the parallel pipeline, but not
enough memory for a lot of detail. But, a lot more could have been done in the 1970's
using thee techniques. The Atari 7800 chipset was indeed, using a lot of other techniques that could have
been invented and used earlier on. The way things could have been.