I still believe that VMS was the superior operating system. And their support was outstanding.
[I remember asking an MS exec what they would do about Linux and their view, apparently for too long, was that since you had to be a stud to use it, they were content to compete with a KISS-like UX. They didLose the entire server market to Nobody In Particular with that strategy.]
The only party able to make “personal Unix” economically viable tho was NeXT’s Apple and I believe we are witnessing the end of that era now as Apple moves everything to embedded (due to 3nm ARM chips fabricated by TSMC that descended from DEC’s StrongARM embedded processors, also included in the DEC settlement with Intel).
Rest In Peace Ken Olsen.
It is interesting that there are, in hindsight, a couple of specifically identifiable missteps.
DEC wasted about one billion dollars on trying to develop the VAX 9000 mainframe-class VAX, when it was already clear that RISC was the way forwards:
https://en.wikipedia.org/wiki/VAX_9000
Remarkably, it came back from this with the Alpha, which shows what it could have done.
If it had instead spent that vast R&D budget on PRISM and MICA, it would have had what became NT to itself.
https://en.wikipedia.org/wiki/DEC_PRISM
https://en.wikipedia.org/wiki/DEC_MICA
PRISM was cancelled; some of the tech was salvaged to make Alpha.
MICA was cancelled; the project lead quit, and took his core team to MS, where they rescued OS/2 3.0 and made it into Windows NT, which has strong resemblances to the planned DEC MICA.
What didKill DEC was the blue blood Board, which had FOMO due to the rise of Microsoft (which, at that time, was still running on the anemic 286/386/486 chips, but still killing it due to NetWare and NETBIOS) which was being led by "one of their own" (sort'a --the estranged son of Rockafeller's lawyer) after having gotten started with just $100K (a kiss goodbye gift after he had dropped out of college?) The little misstep with DEC's first generation of PCs, which were Ken's personal babies, caused him to lose the support of the "desperate" Board and, well, you saw what they did...and One Thing Lead to Another...until:
I wasn't talking about the VAX 8000, but the 9000. Not the same machine, or processor.
> What didKill DEC was the blue blood Board,
Your weird capitalisation here is confusing me. What is "the blue blood Board"?
> the anemic 286/386/486 chips
Hang on, what? That's a timspan from the 1980s to the 1990s, from sluggish 16-bits to competitive 32-bit 2nd gen chips that started to obliterate the competition. It's ludicrous to lump all those together.
> but still killing it due to NetWare and NETBIOS
Not sure what "killing it" means in this context.
Not sure why you'd glue Netware and NetBIOS together.
Not sure what you're saying at all, in fact.
> which was being led by "one of their own" (sort'a --the estranged son of Rockafeller's lawyer)
So many weird unclarified allusions! Who? What? Where?
Do you want to try again without being cryptic and hinting at things that are not really meaningful to anyone else?
RE: "weird capitalization" - "Board" was capitalized because I was referencing a specific board of directors. "Blue blood" was not capitalized because I was characterizing the members of said board. The relevance of making that characterization was to emphasize that those board members were under INTENSE pressure from their peers/ancestors/descendants to achieve at least the level of success that they were all observing coming from Redmond, especially the younger members of their dynasties. I think they were fine with a Pinto-driving living legend like Ken and his armada of actual geniuses (not Geniuses) until Redmond got going. Then FOMO didn't even begin to describe what they were experiencing as "some geeky guy" made hundreds of billions of dollars and became the wealthiest man on Earth, right under their noses. (They've ensured that can never happen again, but I get my account here Closed if I tell how they did it.)
RE: 286/386/486 was the end of the original Intel X86 line (without Alpha-infused technology). They were in various stages of being knocked off by Chinese companies like VIA. Microsoft was porting NT away from that ISA (onto MIPS). Pentium came out of left field and scaled like crazy. I'm not sure if/when it's fair to stop attributing Alpha technology to the success of X86.
RE: "killing it" --NT on even a Pentium III was no match for a VAXcluster. A network of NT boxes interconnected by "field technicians" into a LAN was. (A "field technician" of that era would have trouble interconnecting VAXes --but they had more than ample capability to setup NetWare and NETBIOS networks in the most unusual ways and places.) In this way, it can be said that WindowsNT (and the budget it gained from legacy NetWare installations) literally connected the VAX to death, not unlike what VAX had been doing to IBM mainframes (and their clones) for sometime.
> weird un-clarified Well, you are a good person because when you sense a lack of context, you don't instantly hit downvote (unlike most on Hacker News).
I suggest amplifying this with examples, citations and links, and sticking it in a blog post somewhere that people can comment on. OSnews springs to mind.
It seems to me that you know a lot of the history here that I don't, and I'm a journo working in this field and a former VAX sysadmin. This stuff is on the point of being lost to history now. It NEEDS to be recorded.
And you seem to be a good person to do it.
Don't leave it as a few hints in HN comments.
Even this response explains a little, but it's a tease: it has no citations, no links, no pointers, and no explanations of anything except my specific questions.
You really need to amplify this with 10x more background for a general audience.
But PLEASE do so. It needs to be recorded.
It’s not clear to me whether they would have gotten as much out of it as Microsoft did out of NT, though. The marketing of NT4 (at least) reached down to fairly small server setups (even if wasn’t particularly popular on the workstation side—hardware requirements?), what with the BackOffice branding and everything, and Microsoft wouldn’t have blown you off if you just wanted to give them $5k for a few licenses. As I’m reading it, DEC’s mindset couldn’t really support this kind of strategy.
Also, nearly nobody programs against the NT API as opposed to the Win32 API and things built on top of it. I can’t tell what the intention of MICA was—keep the nice native API essentially private and only expose the VMS “personality” running on top of it, like Microsoft did, or something else?
Oh, absolutely agreed.
MS did very well and the core of NT is a superb product. But merely having a superb product is not enough.
It's hard to see how big and how influential DEC was today. Now, young techies assume that an 8-bit byte is a universal thing, that it's natural and obvious that computers just obviously use binary and that means word lengths that are powers of 2...
Whereas DEC sold vastly successful minicomputers with word lengths of 12, 18, 24 and 36 bits as well as 8-bit and 16-bit machines!
https://en.wikipedia.org/wiki/Programmed_Data_Processor
And those machines inspired the designs and the command lines of CP/M and MS-DOS, and as such, directly and profoundly influenced the PC industry. These were not niche machines, they were the mainstream that led to the standard OS that made Intel a success with the 8080, and that in turn led to the Zilog Z80 that spawned many multi-million-selling 1980s home computers... but it also led to the PC.
DEC OS/8 (and its big sibling TOPS-10) are what led to not only the CP/M and DOS command lines, but also things like 3-letter file extensions (after a dot). Every .TXT and .JPG and .HTM file exists because of the PDP-8.
https://liam-on-linux.livejournal.com/81341.html
And that was a 12-bit machine, not 8-bit, or 16-bit like the PDP-11, the machine that led to the 32-bit VAX, whose VMS is the direct ancestor of MICA and NT.
UNIX was developed for the PDP-7, an 18-bit machine.
The machines and their OSes are nearly forgotten now, and DEC is a footnote, but the weird thing is that both the CP/M–MS-DOS–OS/2–Windows–Windows NT family evolved from DEC OSes on DEC machines, but also the Unix family was designed and implemented by a 3rd party on DEC hardware, and then ported to different DEC hardware.
The two big rival lines of OSes today both can be traced back to the same company's products.
Unix and microcomputers are what killed DEC... It foolishly bet against them, it bet on big iron and tried to turn its most successful minicomputer into a mainframe.
But even despite that, and wasting a mind-blowing amount of money, it still came back and produced the first 64-bit RISC chip, and the first 64-bit Unix to run on it, and it got a native version of the industry-dominating OS which grew out of its own cancelled project. And that chip was also the target of the first ever port of Linux to something non-x86.
It was an incredible comeback. It's tragic that ultimately it wasn't enough.
It would have been interesting if someone had built an open-source VMS-like OS instead of Linux.
In what concerns the kernel, userspace is another matter.