OT: Why it is better to pretend you know nothing about computers

It's true that one of the original architects of Windows NT was previously a senior architect of VMS. No doubt he took some of the good ideas with him, and early on in Windows versions, it was possible to see similar concepts as there are in VMS. However Windows never (still hasn't...) got anywhere near the reliability of VMS - the usual joke inside digital was when VMS had to reengineer the up time display to allow more digits, because there were systems that had not been rebooted for several years, I think the record was something like sixteen years...

Later it was discovered that MS had "accquired" the source code to clustering from VMS, and although digital cut a deal with MS (rather than suing their a****s off !!) so that MS could legally use the algorithms, Windows has still not got a clustering system that remotely compares with the almost transparent way that clusters work in VMS.

Graham (Seventeen years at digital/Compaq/hp as a VMS application developer, now gone over to the dark side :-) )

Reply to
GAP
Loading thread data ...

I've looked at the source code of both and there are definite similarities in places. And NFTS is an obvious descendant of ODS-2. Not to mention that VMS and WinNT shared the same lead architect - and quite a bit of the team.

A friend of mine worked on SMP code for an operating system kernel I was involved in, back in the early 1980s (I did the SMP specifics for a particular mainframe). He did some of the SMP code for VMS 5.x, and moved on to Microsoft with the others.

Reply to
Bob Eager

The structure was what I meant. Source code would have been pointless, as most of the VMS kernel was written in assembler (I've seen it, and have a copy somewhere).

Take a look at ODS-2 and NTFS for a good example.

Reply to
Bob Eager

Yeah, but what happened to file version numbers along the way?

Reply to
Andy Burns

I suspect with a little shim work you could use NTFS streams to emulate file versioning. Performance would be lousy, though - unintelligent copy-on-write.

Cheers - Jaimie

Reply to
Jaimie Vandenbergh

The code was not borrowed as such (VMS was written in VAX assembler for starters), but there are many architectural similarities.

Some are detailed here:

formatting link

Reply to
John Rumm

Where did I say I don't like Windows? Not having a use for *nix or any other o/s, its all I use.

S'ok, I won't try humour in here again.

Reply to
SteveH

from Wikipedia: "Microsoft hired a group of developers from Digital Equipment Corporation led by Dave Cutler to build Windows NT, and many elements of the design reflect earlier DEC experience with Cutler's VMS and RSX-11." and you'll find the full story in "Show Stopper!: The Breakneck Race to Create Windows NT and the Next Generation at Microsoft" - a very good read.

Reply to
Tony Bryer

I think it's more now...although it's cluster uptime, rather than a single machine. Windows of course is much bigger - the relatively small team (and the ethos) at Digital was a great help. But yes, the Microsft 'way' is never going to result in something that reliable.

I still have three VAX machines at home...!

Reply to
Bob Eager

A lot of people have an inbuilt hatred of them, so I suspect it was politics. And trying to keep things 'simple'...I can imagine uncontrolled version numbers causing full disks.

I wish they were there too...

Reply to
Bob Eager

Poor chap then had to add the Windows drive letter rubbish on top of NT to make it like the other versions, as I understand it.

Reply to
Tim Streater

The monitoring app I wrote in the 80s for a MicroVAX was typically up for 9 months at a time. Only got restarted because of the annual site-wide power outage.

Reply to
Tim Streater

Which models? And are they on?

Reply to
Tim Streater

The interesting thing is the origin of the drive letters. Obviously back to MS-DOS, before that CP/M, and you get right back to 1970 and OS/8 on the PDP-8, which had drive names (usually 3 letters, like SYS: and DSK:). OS/8 begat other systems such as DOS/BATCH and RT-11 (PDP-11) whch had the same thing, and they begat VMS. VMS had drive names too (typically DUA0:, RXA3: etc.), so using a single letter wouldn't have been a very big step...

Reply to
Bob Eager

That's an awful lot of microfortnights! (VMS joke...)

Reply to
Bob Eager

MicroVAX 2000. Which isn't. VAXStation 3100 (which is currently at work for a history lecture I'm giving). VAXStation 4000 Model 60, which is in use, but not on all the time due to noise and power considerations.

I know they're no big ones, but the 4000 has more VUPs than one CPU of our old 8800.

Then there's the 11/23, and the two replica PDP-8s....

Reply to
Bob Eager

VAXserver 3300 here, which was only interested in booting from TK70 last time I tried, VMS5.x and 6.x installers wouldn't recognise disc controller AFAIR.

Reply to
Andy Burns

Except it was easy to set the version limit on a directory tree - York had it at 4, but it was user changable.

2 would give a unix-y experience and so much better that .bak .old ~ and whatever other random extensions the app write dreamt up lying about :)
Reply to
Tim Watts

And still didn't implement a direct equivalent of logical search drives in VMS which were a) a trivial algorithm, b) extremely useful and would have made a lot of sense with Windows .ini files and other config at the time (and would make a lot of sense with unix too).

Reply to
Tim Watts

I agree. We had it set at three - my decision. I didn't say it was a

*rational* hatred...but it came from our earlier ICL VME system where there were no version limites - and, incredibly, no PURGE command!
Reply to
Bob Eager

HomeOwnersHub website is not affiliated with any of the manufacturers or service providers discussed here. All logos and trade names are the property of their respective owners.