Re: [TUHS] Windows roots and Unix influence (was Re: Happy birthday, Ken Thompson!)

On 2/4/18, Wesley Parish <wobblygong(a)gmail.com> wrote: By the mid-to-late 1980s, advances in CPU technology tilted the performance field in favor of RISC architectures vs. complicated CISC such as VAX or x86. Both DEC and Intel were looking for alternatives. Intel eventually settled on a VLIW architecture that became Itanium. DEC's first attempt at a VAX successor was a RISC architecture called PRISM. It was developed by Dave Cutler's team in Seattle. On the software side, they were working on an OS called MICA. It was to be a successor to the VAXeln microkernel-based realtime system and was to have both VMS and Unix kernel interfaces (personality modules). Someone (Gordon Bell, I think) once said of DEC's decision making process, "any decision worth making once is worth making again". In typical DEC fashion, another RISC architecture was designed by DEC's east coast engineering team. This architecture eventually ended up being called Alpha. The east coast won the political battle. PRISM and MICA were cancelled. The idea of a single kernel with multiple OS personalities was dropped; VMS and Unix were ported separately to Alpha. After PRISM was cancelled, Dave Cutler left DEC and went to Microsoft as architect for their new OS. Under the covers, the original Windows NT looked a lot like MICA, VAXeln, and VMS before it. Not surprising since they shared the same designer. Like MICA, Windows NT was microkernel-based, with OS personality modules layered on top. There were two of these originally: Win32 and POSIX. Microsoft and Intel were having a little lover's quarrel at the time--Intel didn't like Microsoft's forays into the hardware side of things with Xbox and the like; Microsoft wasn't pleased by Intel doing its own compilers, etc. This led to Microsoft porting Windows NT to both PowerPC and Alpha. Neither port caught on in the marketplace. -Paul W.