RichardC (tich.delete@this.pobox.com) on April 9, 2017 7:54 am wrote:
> Linus Torvalds (torvalds.delete@this.linux-foundation.org) on April 8, 2017 11:43 am wrote:
>
> > MIPS was definitely used inside of DEC, and alpha was supposed to be the replacement. In many ways,
> > alpha was "MIPS done right" - get rid of all the broken parts, and extend it to 64 bits cleanly.
>
> AFAIK DEC's MIPS workstations were only running Unix, when their business need
> was for a faster and/or better price-performance way to run VMS and VMS apps
> (ideally without even recompiling VAX/VMS binaries ...). So in that sense DEC's
> tactical use of MIPS was barely dipping their toe into the water.
>
> Running VMS was a tough requirement. VMS was developed in 1975-77, for VAX-11. My impression
> has always been that it was written almost entirely in VAX assembler: at the very least,
> it had so much VAX-specific code that the later porting to Alpha and then IA64 involved
> completely separate source code: one tree for VAX/VMS, one tree for OpenVMS (with small
> variants of that for the Alpha and IA64).
>

This is what I don't understand, is how Digital Equipment Corporation - had inherited not only 'one' but both of the major existing branches of operating system software philosophy - and had both of them inside of it's 'stable' at one stage in the 1970's, . . . and they still couldn't manage to make a horse-race out of it.

http://www.realworldtech.com/forum/?threadid=166772&curpostid=166824

Bob Metcalfe, has an oral history (bearing in mind, that someone involved with ARPA net, and Ethernet later at Xerox Parc, would have had many a run in, with the pre-1984 AT&T company).

http://www.realworldtech.com/forum/?threadid=166772&curpostid=167066

One thing he said in the oral history, I laughed at. What did AT&T ever do, except for inventing the transistor, Unix and C-language? It's like the Monthy Python question: What did the Romans ever do for us? I would have rated those together, as pretty good contributions. His point though, was that real research only gets paid for, when there is a 'monopoly'. However, monopolies while being able to fund research, are also reluctant to drive new things, in the way that they could be. He argued therefore, for the model such as the government funded research carried out, with the ARPA-net that he was involved with.