38 ms·

They never understood how to build a product line or a roadmap. To them, the Amiga was a on/off kind of thing, a product they acquired once and planned to sell

They never understood how to build a product line or a roadmap. To them, the Amiga was a on/off kind of thing, a product they acquired once and planned to sell for some years. They would have happily sold some new, completely incompatible home computer, afterwards.

vidarh 3y ago
The fascinating thing is that they were sort-of understanding they couldn't really do that anymore towards the end, but too late. Hombre, their last attempt, was one one hand meant to provide a clean break by moving to a PA-RISC core (chosen to allow for Windows NT compatibility but in retrospect might just have deferred their death) in a dual-chip pair that'd combine the CPU, DMA, an upgraded blitter that was closer to a basic GPU (transforms, texture mapping, gouraud shading, z-buffering), a 64-bit RISC-like copper replacement, 16-bit sound, with a display controller on the other chip. But to avoid breaking backwards compatibility, they also planned putting an entire AGA Amiga on a chip, and recompile AmigaOS, so they could end up with sort-of reverse of what we got with the PowerPC expansion boards or the older Janus Bridgeboards (putting an x86 CPU straddling both a Zorro and ISA buses, letting you run PC software at full speed in a window on the Amiga), of having the new CPU run the OS, and using the embedded 68k Amiga-on-a-chip to run legacy software. A valiant attempt, but of course way too late - Commodore ran out of money before completing even the chipset before Hombre in their pipeline. Hombre would've been a big break, though, with what looks like they had something resembling an actual plan for a very broad range of uses from the start instead of their usual haphazard Darwinist "survival of whatever nearly ready design still happens to please the right executives" Commodore usually practices. It ranged from putting it on a high end graphics card, to set-top boxes, CD32 like entertainment systems, all the way up to high-end machines running AmigaOS and WindowsNT and based around a machine concept from David Haynie (Acutiator, going back to '91) that planned for exchangable modules for everything including the CPU, with the ability to even adapt the bus control account for different CPU requirements. It's an agonisingly fascinating "what if" of alternative hardware, while at the same time in retrospect having some serious flaws (PA-RISC...) and probably being way too ambitious for Commodore even at its peak given the amount of discipline shepherding a project that ambitious over the line without some executive ego trip derailing it would've taken...

vidarh 3y ago

The fascinating thing is that they were sort-of understanding they couldn't really do that anymore towards the end, but too late. Hombre, their last attempt, was one one hand meant to provide a clean break by moving to a PA-RISC core (chosen to allow for Windows NT compatibility but in retrospect might just have deferred their death) in a dual-chip pair that'd combine the CPU, DMA, an upgraded blitter that was closer to a basic GPU (transforms, texture mapping, gouraud shading, z-buffering), a 64-bit RISC-like copper replacement, 16-bit sound, with a display controller on the other chip. But to avoid breaking backwards compatibility, they also planned putting an entire AGA Amiga on a chip, and recompile AmigaOS, so they could end up with sort-of reverse of what we got with the PowerPC expansion boards or the older Janus Bridgeboards (putting an x86 CPU straddling both a Zorro and ISA buses, letting you run PC software at full speed in a window on the Amiga), of having the new CPU run the OS, and using the embedded 68k Amiga-on-a-chip to run legacy software. A valiant attempt, but of course way too late - Commodore ran out of money before completing even the chipset before Hombre in their pipeline. Hombre would've been a big break, though, with what looks like they had something resembling an actual plan for a very broad range of uses from the start instead of their usual haphazard Darwinist "survival of whatever nearly ready design still happens to please the right executives" Commodore usually practices. It ranged from putting it on a high end graphics card, to set-top boxes, CD32 like entertainment systems, all the way up to high-end machines running AmigaOS and WindowsNT and based around a machine concept from David Haynie (Acutiator, going back to '91) that planned for exchangable modules for everything including the CPU, with the ability to even adapt the bus control account for different CPU requirements. It's an agonisingly fascinating "what if" of alternative hardware, while at the same time in retrospect having some serious flaws (PA-RISC...) and probably being way too ambitious for Commodore even at its peak given the amount of discipline shepherding a project that ambitious over the line without some executive ego trip derailing it would've taken...