Ask HN: Could Commodore have survived with a GEOS-powered C128 and no Amiga?
And after the GEOS-powered C128, it could have then released a GEOS-powered C1000 running on a 68000 or Intel chip with software emulation of C64 and C128 software.
7 comments
GEOS 2.0 was in fact released for the C128 in 1989. It did not spur continued interest in the machine, similar to how GEOS 1.x was not a primary driver for interest in the C64.
GEOS was also released for the Apple II and PC/GEOS for IBM PC compatibles in 1988 and 1990 respectively, even beating Windows 3.0 to the market. Neither version established any substantial market share.
There simply was never substantial demand for GEOS. While actual sales figures are hard to come by, it seems very likely there was substantially less interest in GEOS than in the Amiga platform. It would not have "saved" Commodore.
GEOS was also released for the Apple II and PC/GEOS for IBM PC compatibles in 1988 and 1990 respectively, even beating Windows 3.0 to the market. Neither version established any substantial market share.
There simply was never substantial demand for GEOS. While actual sales figures are hard to come by, it seems very likely there was substantially less interest in GEOS than in the Amiga platform. It would not have "saved" Commodore.
But it could have provided a unified interface GUI on the C64, C128, and future models possibly running on different CPUs.
Commodore's strategy was to sell an entry-level computer as cheaply as possible, whereas the Amiga's initial selling point is that it was cutting-edge but still somewhat affordable. If you want to see what Commodore's answer to the Amiga would have been, look at the Atari ST: cheaper, still using the 68000, but lacking the advanced features which made the Amiga special.
Commode would have needed a real path to the future.
One problem with the 6502 was the tiny 64k address space, the other was that the 6502 was a terrible compiler target with the result that compiled languages for the 6502 usually used virtual machine techniques that gave awful performance like the atrocious UCSD p-System.
There was the 65816 which was clocked higher and had a bigger address space but did nothing for the compiler problem and did not have 24 bit index registers to go with the bigger address space and didn’t have anything like the segments in the 8088/86 that let you do pretty well despite not having full size index registers. In an alternate universe there could have been a path to 24 bits (80286) and then 32 bits (80386) that was compatible with the 6502 but there wasn’t.
The Apple //gs was an impressive machine that looked good compared to the very expensive Mac 2 and I think that’s what your ‘Super 128’ might have been at best. Have you seen
https://www.commanderx16.com/
? I also think the only 24-bit extension of an classic CPU that I like is
https://en.wikipedia.org/wiki/Zilog_eZ80
which really has 24 bit arithmetic and index registers and is based on an architecture which is compiler friendly.
One problem with the 6502 was the tiny 64k address space, the other was that the 6502 was a terrible compiler target with the result that compiled languages for the 6502 usually used virtual machine techniques that gave awful performance like the atrocious UCSD p-System.
There was the 65816 which was clocked higher and had a bigger address space but did nothing for the compiler problem and did not have 24 bit index registers to go with the bigger address space and didn’t have anything like the segments in the 8088/86 that let you do pretty well despite not having full size index registers. In an alternate universe there could have been a path to 24 bits (80286) and then 32 bits (80386) that was compatible with the 6502 but there wasn’t.
The Apple //gs was an impressive machine that looked good compared to the very expensive Mac 2 and I think that’s what your ‘Super 128’ might have been at best. Have you seen
https://www.commanderx16.com/
? I also think the only 24-bit extension of an classic CPU that I like is
https://en.wikipedia.org/wiki/Zilog_eZ80
which really has 24 bit arithmetic and index registers and is based on an architecture which is compiler friendly.
For some reason there's something quite comforting to me knowing that TI still sells calculators with Z80-based chips not much different than what was in the TRS-80 from 1977. I think I like that fact quite a lot more than knowing that my phone is a lot more powerful than a desktop computer was when Windows XP came out.
It's funny. Circa 1990 I think calculator enthusiasts were much more likely to be into HP, particularly devices like
https://en.wikipedia.org/wiki/HP_48_series
but TI really took over most of the market and has most of the enthusiast interest. I have an HP Prime but I think I'm the only one, it's a really nice calculator but the algebraic mode is better than the RPN mode which makes not feel very 'HP' and at the core it is another ARM device.
https://en.wikipedia.org/wiki/HP_48_series
but TI really took over most of the market and has most of the enthusiast interest. I have an HP Prime but I think I'm the only one, it's a really nice calculator but the algebraic mode is better than the RPN mode which makes not feel very 'HP' and at the core it is another ARM device.
Maybe if the GEOS-powered C128 also had a 65c816 CPU, a vastly more powerful CPU than the 8502 without being as expensive as the 68000 was at the time. The C128 needed more memory.
Commodore needed more memory. We already had megabyte machines in the late 80's.
Commodore needed more memory. We already had megabyte machines in the late 80's.