4 ms·

I'm not sure I would describe a team that releases an OS without memory protection, pre-emptive multitasking, or a 32-bit API several years after the introducti

I'm not sure I would describe a team that releases an OS without memory protection, pre-emptive multitasking, or a 32-bit API several years after the introduction of the 386 a "heck of a systems programming shop". The NT kernel, developed around that period, was good for its time though. Windows 3.x ... not so much.

UnoriginalGuy 13y ago
I hate retrospective analysis of computing history which try to criticise things without understanding the context/reasoning, and instead look at it in a modern lens (e.g. people that call Windows 95 "bad"). You just contrasted Windows NT (1993) with Windows 3.0 (1990). Apart from the fact that NT is a MUCH newer OS released later (with 3.1/3.11 being upgrades onto an earlier OS), the system requirements of the two are quite starkly different. For one example, Windows 3.0 runs on 384K of RAM. Windows NT 3.1 runs on 12 megabytes. Windows 3.0 also runs on 8 and 16 bit CPUs, while NT is 32 bit only. If they had released Windows 3.0 as a 32 bit OS Microsoft literally might not exist as they do today, since Windows 3.0 was targeting cheaper IBM clones which were very behind the curve technologically. 32 bit would also have meant perhaps doubling the minimum memory requirement of the system. In general I think very little of your post.
- brokenparser 13y ago
  Windows 3.0 also runs on 8 and 16 bit CPUs, while NT is 32 bit only. There isn't a version of Windows for any 8-bit CPU. It did, however, run on an Intel 8088 which is a 16-bit CPU with an 8-bit external data bus. (Until Windows 3.1 came along which dropped real address mode support.)
- mcv 13y ago
  This is not a modern lens. People back then also called it bad. I remember when Windows 95 was released, people mockingly called it MacOS 89, RiscOS 88, or AmigaOS 85.
  - andyjohnson0 13y ago
    I wonder who these "people" were. In 1990 I was using Apollo and HP workstations that made win3 look like a toy, but these were extremely expensive, high-spec devices. People who were accustomed to such systems may well have "called it bad". But for those with ordinary PCs at that time, windows 3 was perceived very differently.
  - snori74 13y ago
    People on Mac, RiscOS or AmigaOS may well have done so, but by 1995 two of those three were dead. For PC users, Win 3.11 and Win 95 were both BIG advances at the time, no matter how they look to us in retrospect.
    - new299 13y ago
      MacOS and RiscOS at least were still going in 1995. Not to mention OS/2 was around. There really were better alternatives to both 3.11 and 95. A shell sitting on top of DOS was always quite hacky. They were big advances for people in the IBM compatible ecosystems who needed DOS compatibility, but not in a general sense.
- vidarh 13y ago
  Consider systems of the same era: A 1985 Amiga had pre-emptive multi-tasking on a system with a 16 bit data bus (though 32 bit internally), with 512K (technically 256K, but really it was not realistic to use a system like that very well). By 1990 there were plenty of networking solutions. There were also virtual memory extensions (but not memory protection). It was bad for the era. It did well because the competing solutions only ran on expensive single-source hardware and/or were not seen as serious enough (even the big-box Amiga's were, aside from cost, negatively affected by being connected to the "game machines" on the low end).
caf 13y ago
WIN32S implemented a subset of the 32-bit Windows NT API on Windows 3.11.
andyjohnson0 13y ago
Unfair. Windows 3.x had to be able to run on an 8086 (real mode), because that was what most people had. It could also make use of a 286 (standard mode) or 386 (enhanced mode) if available. Given the hardware constraints I think it was an impressive piece of work for the time.

UnoriginalGuy 13y ago

I hate retrospective analysis of computing history which try to criticise things without understanding the context/reasoning, and instead look at it in a modern lens (e.g. people that call Windows 95 "bad"). You just contrasted Windows NT (1993) with Windows 3.0 (1990). Apart from the fact that NT is a MUCH newer OS released later (with 3.1/3.11 being upgrades onto an earlier OS), the system requirements of the two are quite starkly different. For one example, Windows 3.0 runs on 384K of RAM. Windows NT 3.1 runs on 12 megabytes. Windows 3.0 also runs on 8 and 16 bit CPUs, while NT is 32 bit only. If they had released Windows 3.0 as a 32 bit OS Microsoft literally might not exist as they do today, since Windows 3.0 was targeting cheaper IBM clones which were very behind the curve technologically. 32 bit would also have meant perhaps doubling the minimum memory requirement of the system. In general I think very little of your post.

brokenparser 13y ago
Windows 3.0 also runs on 8 and 16 bit CPUs, while NT is 32 bit only. There isn't a version of Windows for any 8-bit CPU. It did, however, run on an Intel 8088 which is a 16-bit CPU with an 8-bit external data bus. (Until Windows 3.1 came along which dropped real address mode support.)
mcv 13y ago
This is not a modern lens. People back then also called it bad. I remember when Windows 95 was released, people mockingly called it MacOS 89, RiscOS 88, or AmigaOS 85.
- andyjohnson0 13y ago
  I wonder who these "people" were. In 1990 I was using Apollo and HP workstations that made win3 look like a toy, but these were extremely expensive, high-spec devices. People who were accustomed to such systems may well have "called it bad". But for those with ordinary PCs at that time, windows 3 was perceived very differently.
- snori74 13y ago
  People on Mac, RiscOS or AmigaOS may well have done so, but by 1995 two of those three were dead. For PC users, Win 3.11 and Win 95 were both BIG advances at the time, no matter how they look to us in retrospect.
  - new299 13y ago
    MacOS and RiscOS at least were still going in 1995. Not to mention OS/2 was around. There really were better alternatives to both 3.11 and 95. A shell sitting on top of DOS was always quite hacky. They were big advances for people in the IBM compatible ecosystems who needed DOS compatibility, but not in a general sense.
vidarh 13y ago
Consider systems of the same era: A 1985 Amiga had pre-emptive multi-tasking on a system with a 16 bit data bus (though 32 bit internally), with 512K (technically 256K, but really it was not realistic to use a system like that very well). By 1990 there were plenty of networking solutions. There were also virtual memory extensions (but not memory protection). It was bad for the era. It did well because the competing solutions only ran on expensive single-source hardware and/or were not seen as serious enough (even the big-box Amiga's were, aside from cost, negatively affected by being connected to the "game machines" on the low end).

caf 13y ago

WIN32S implemented a subset of the 32-bit Windows NT API on Windows 3.11.

andyjohnson0 13y ago

Unfair. Windows 3.x had to be able to run on an 8086 (real mode), because that was what most people had. It could also make use of a 286 (standard mode) or 386 (enhanced mode) if available. Given the hardware constraints I think it was an impressive piece of work for the time.