Comment by panick21_ | Hacker News Reader

Comment by panick21_ | Hacker News Reader

Parent Full thread

The more I learn about history of computing, the more it seems like DEC canceling of Project PRISM (and Project MICA) is at the center of a lot of history.

Project MICA would have yield a very modern OS that could have supported both Unix and VMS interfaces and it would be only natural to add whatever Windows would come up with in addition to that. That PRISM and MICA seem to be a real paths forward for everything from workstation to multicore servers. We later saw with Alpha what DEC was capable of and PRISM also did a lot of compiler work that was also used for the Alpha compiler later. But that was quite a bit later of course.

First of all, Cutler and lots of others went to Microsoft and developed Windows NT. Without Cutler, this project would look very different. Bill Gates was seriously arguing for having no virtual memory on this OS! Absolutely crazy to think what would have happened.

If DEC had expanded, rather then contracted its team there, it would have been quite a difference, with two major companies developing a next generation OS in the area. I think DEC was generally more popular for technical people then Microsoft.

Then for DEC, their workstation strategy went to MIPS/Ultrix witch was moderately successful but that whole thing was then killed in favor of Alpha/OSF1. This created a lot of bad blood by the people that were stuck on MIPS/Ultrix. DEC was one of the few companies that really went all in on the whole OSF stuff. Just like Sun, it might have been smarter to stick with BSD base rather then System V SVR4 base (love to hear from somebody inside about this topic).

Because PRISM didn't happen over the next couple years the VAX line suffered from pretty bad price to performance. Management also seem not committed to VMS as OSF and NT became big parts of their strategy.

So this post about running Windows NT+Alpha is really kind of a result of this canceling of PRISM/MICA.

I reading some books on DEC now but during some online browsing I found this interesting document from somebody at DEC from late 1994: https://www.dgregscott.com/wp-content/uploads/2020/11/demise...

Gorden Bell's Post Script on 'DEC is Dead, Long Live DEC' was also a great read.

>Bill Gates was seriously arguing for having no virtual memory on this OS!

That’d surprise me. This was a time when one of the major limiting factors to NT adoption was it needing “a lot” of RAM. And he’d had plenty of experience with virtual memory, including his own company’s Xenix and OS/2.

Had DEC no cancelled, NT would have never been, and Gates wouldn't have had the confidence to abandon OS/2.

So in the end we all won, just Dec had to die.

But DEC only saw Altavista as a toy, something for showing off big systems, or doing personal desktop search. The missed such incredible opportunity.

Not sure why you say windows not doing OS/2 is so great?

OS/2 was a reasonable good OS and a history based on OS/2 rather then Windows 95 and friends seems to be a pretty good path forward.

And OS/2 being Microsoft and IBM would have made Microsoft not quite as powerful and dominate.

> The missed such incredible opportunity.

DEC was sold for around 10 billion, just a few years later Altavista alone was sold for around 2 billion.

And Altavista was valued basically 0 in the Compaq acquisition. Compaq seem to really only value the Services division, everything else they didn't care about.

My school had at least a thousand DEC MIPS/Ultra/pmax workstations in 1993. When they announced they were killing the line we bought thousands of Suns and HP-UX machines. I remember seeing 2 Alpha machines in a special computer lab and I could log in but none of our engineering software ran on it. Maybe in a few years but I graduated by then.

This is kind of the crux of the issue. Alpha came out with great performance (but pretty expensive) in general but some issues related to performance of Unix and strings.

They killed the MIPS/Ultrix line but gave people no path to move to Alpha/Ultrix. Ultrix was never ported to Alpha. Only Alpha/OSF1 so a fundamental hardware and software shift.

Its easier to go to SPARC/SunOS or SPARC/Solaris and they had lots of machines from cheaper to more expensive. For schools they often bought the slightly cheaper workstations.

Alpha 64bit wasn't really a hugely important feature that point in time. This again where starting with 32-bit PRISM would have been a better move.

At the same time, their VAX line was still not Alpha, they bet on NVAX for quite a long time. So the VMS crowed didn't need to switch to Alpha.

DEC wanted to make Alpha an industry standard but then they had little interest in working with Apple to get volume. The big workstation/server vendors had their own ISAs and since DEC also wanted to compete in those market simply adopting a new standard from DEC was just not gone happen. And as a new standard for lower priced machines it just not gone happen, 64-bit was totally wrong for markets like that and you would need Windows/Intel to even have a shot at that. It was also not good for embedded so DEC went with StrongARM.

DEC just made really bad strategy choices starting in late 1988s. By 1986 they had started the VAX9000 project. That might be fine but by 1988 you have to realize that project was dead and simply kill it, and rush out VAX based on PRISM with nice high quality VMS port.

Reply on news.ycombinator.com