Public Access OpenVMS System | Hacker News Reader

Public Access OpenVMS System

decuserve.org

102 pointsBSDobelix53 comments

VMS was the first multiuser system I used in 1991. I was amazed that a single computer could have 300 users logged in all doing their own thing.

I learned to program in C and tried using Borland Turbo C on a DOS. Go outside of the array index and you could easily crash the entire machine which meant reboot and try it again. I thought the concept of memory protection was incredible! Why didn't every computer have that?

It was the first time I saw email, sending messages, and the phone utility for chatting with other users while waiting for your programs to run in the background.

I moved on to an IBM RT running AIX and got hooked on Unix. That's the system I wanted and have been running Linux since 1994. But VMS still holds a special place.

On of my favourite hardware investments as a teenager trying to learn c and assembly under dos was a caching hard drive controller.

It made those post crash reboots so much faster!

Talk about solving the wrong problem...

My roommate with the 386 running DOS tried to explain extended vs. expanded memory and then near and far pointers in C to me. It didn't make any sense to me. I just asked him "Why isn't memory just memory like on the VAX?" He was so used to the limitations of 16-bit x86 and DOS. Neither of us even understood the concept of a "32-bit flat memory model."

I had been spoiled by the VAX and didn't know it was a $200,000 or $1 million system. I found this price list from 1991 but it's hard to figure out how much a system cost.

https://bitsavers.trailing-edge.com/pdf/dec/priceLists/US_Sy...

VAXstations and MicroVAXes weren't that expensive. I knew a guy with a small business that had a VAXStation 3100 or maybe it was a 4000, back in 1991 or 92.

I think it was a VAXserver 3000 series. The pricelist I found shows it was probably in the $15K to $25K range

I have a MicroVAX in my living room that I use as an end table.

Rebooting on the early PC was so common, they had a hotkey for it—Ctrl+Alt+Delete.

David Bradley, an engineer who worked on the original IBM PC, invented the combination which was originally designed to reboot a PC. "I may have invented it, but Bill made it famous."

VMS had asynchronous programming to its core. Most system calls could be performed synchronously or async. You just had to give the async call a flag that would indicate when the response was ready. So you could read a file or connect to the network, continue execution and when you actually needed the answer, just sys$wait(flag). Or have callback functions. It worked extremely well as some else mentioned, 64Mb of RAM was enough to handle 200 interactive users with RDB (database IMO far superior to Oracle that was acquired by them and promptly buried)

> VMS had asynchronous programming to its core. Most system calls could be performed synchronously or async.

This continued, of course, in its spiritual successor Windows NT.

Yes, same architect (Dave Cutler), and WNT is (V+1)(M+1)(S+1) though they probably didn't do that on purpose.

> though they probably didn't do that on purpose.

Of course they didn’t (wink wink)

For those curious, look up IO completion ports. Back when all I knew was select(2), it blew my mind.

> This continued, of course, in its spiritual successor Windows NT.

I think NT still had problems with interactivity when the parralel port was used (interrupts)

I can't say I did much w/ NT and parallel ports beyond attaching printers. I'd imagine it could be pretty awful in the "SPP" days, especially with people trying to connect ZIP drives, "Backpack" CD-ROM drives, tape drives, etc. I suspect floppy drives, which I also didn't use under NT very much, could also play hell w/ interactivity.

My first serious (multiuser) system experience was VMS. The concurrency, the power of DCL, the record based nature of files, the mail box concept for process communication, the accounting system, the granular privs, the cluster concept. I came to UNIX second and still miss some of these things.

I still haven’t found anything that matches the power, functionality, and flexibility of VMS batch queues.

For me, the distributed lock manager, queues (SYS$ENC / SYS$DEQ), IPC via message passing (mailboxes), SYS$QIO to literally queue IO requests, event flags (SYS$SETEF / SYS$WAITFR)? Amazing clustering, - UNIX just seemed so backwards after learning VMS first! (Apologies though if I got any of those system call names wrong, its been a long time!)

You can also follow this guide to run your own VMS 1.0 on a VAX 11/780 emulator.

https://gunkies.org/wiki/Installing_VMS_V1.0_on_SIMH

Or your modern OpenVMS V9.2 x86 as a VM:

Community license:

https://vmssoftware.com/products/licenses/

So unfortunate VMS was never open sourced like Solaris was. It was really interesting, and deserved a fate other than obscurity. Would be cool if zOS was open sourced, but on the other hand the various cloud control planes seem like the "real" OS nowadays in many settings.

The original VMS kernel was written in assembler. It is doubtful that anyone would want it.

I don't know how later versions changed.

In any case, Windows NT (re)used a number of concepts from VMS.

https://en.wikipedia.org/wiki/Windows_NT#Development

"Although NT was not an exact clone of Cutler's previous operating systems, DEC engineers almost immediately noticed the internal similarities. Parts of VAX/VMS Internals and Data Structures, published by Digital Press, accurately describe Windows NT internals using VMS terms."

I remember one disk defragmentation utility I used on NTFS, Diskkeeper IIRC, was originally written for VMS.

DEC used to ship source listings to customers. Not in a form you could compile - a CD-ROM full of PDFs. Although not freely redistributable, there are ISO files floating around the Internet

For MVS, the old public domain versions came with some source code, but it doesn’t match up well with the binaries (the supplied source is an older version missing many patches the binaries have). Newer copyrighted versions they made source listings available on microfiche-it would be great if someone who still has those microfiche were to scan them, although I think people are hesitant to do so since they worry IBM might react negatively (they are copyrighted and possibly under NDA too). And then they stopped issuing microfiche, but apparently they replaced it with some system that IBM customers can log in to and look at some of the source code file-by-file but not download it-I imagine they log every file you look at and if you look at too much they might ask you what you are doing. I know other vendors have similar systems-e.g. Oracle has such a system to let their ERP customers look at some of their ERP source code. (Or at least they did when I worked there years ago-I assume they still do.)

MVS and z/OS always had another issue - large chunks of them are written in a proprietary programming language (PL/S or its successor PL/X) for which IBM has never released a compiler. Well, there was a period many years ago when IBM was licensing their PL/X compiler to ISV’s relatively cheaply, but that didn’t last long. Using a proprietary language with no publicly available compiler helped discourage customers from modifying the source code they received, and also helped discourage would-be cloners, especially in the early days when the idea of software copyright was still very new and hadn’t yet been extensively tested in court.

I also have heard they still ship some source for JES2, a bit of an exception to the general rule of “object code only” (OCO).

Indeed, I have a physical CDROM of the Tru64 source. It's interesting to look at.

Have you ever been able to get a community license? I've submitted a request for one several times over the last year or two and never received a response, not even a denial.

Not OP but I got one. Some days ago I got another mail with a new version followed by another mail with a bugfix.

I have one, but I haven’t been able to install it on a KVM machine. I’ll need to dig into the docs to see what am I doing wrong.

Yeah i got one too (around 3 years ago), and some days ago they sent me a email with the updated image.

My request was ignored at first, after i sent a new request with my business email i got a community license link sent within an hour.

Can't take a cosy nap on top of a VAX 11/780 emulator, though.

didn't work with simh 3.12-2 (linked on the page).

Installing simh from "open-simh" and everything works according to instructions, I tried using the version V3.12-5 mentioned on the page https://simh.trailing-edge.com/ and this (and all 3.x?) is the "classic" simh and didn't support the referenced vax780.ini, but the version from github (4.x?) worked. Very cool to see vms boot!

Reminds me of the old Deathrow cluster. Sadly it shut down years ago after a hardware failure.

https://news.ycombinator.com/item?id=8118287

I'm familiar with a site that once had around 30 Vaxen for scientific computing in the late 1980s. This included Vax 11/780s, 11/750s, uVax I's, uVax II's and eventually a Vax 8800 used to feed a Cray Y-MP. All long gone. Linux has moved in and taken over.

Does anyone remember another public VMS system called "Carl" or Carl.best.net or something like that? I had fun playing with VMS on it, but that was years ago.

It looks like even though VMS Software Inc have taken over development of OpenVMS, HP still hosts the v8.4 documentation set, for instance here's the system services reference manual:

http://h41379.www4.hpe.com/doc/84final/4527/4527pro.html#fir...

what interesting software is there to run on a VMS machine?

I believe the point of VMS is to run boring software as reliably as possible. Train-ticketing backends, air-traffic control, warehouse management etc.

It's interesting to see how little effort it takes to keep stuff running.

I used to work for a health informatics company and our department level systems ran on OpenVMS. Radiology management, pharmacy management, etc.

talk? maybe ytalk?

I still miss bitnet.

It's still here (HNET) ;)

https://moshix.dynu.net/

https://github.com/moshix/BITNETServices

Moria (now Angband) was originally on VMS, written in VAX Pascal.

We remember it well because it was distributed as a single, large file, and our 11/730 just choked and died trying to compile it.

The MicroVAX was much better at coping with it.

VAX Notes. Perhaps the original groupware. Not really useful without collaborators any more, since DEC is dead and took that cultural fork with it.

See https://thoughtsofanidlemind.com/2012/12/14/vax-notes/

and the referenced paper

https://thoughtsofanidlemind.com/wp-content/uploads/2012/12/...

VMS-empire ;)

http://www.catb.org/~esr/vms-empire/

I always played Rogue and Trek on the VAX when I had to go in with my mom when she worked weekends.

wooohooo... it had been a long long time since I last entered VMS command in a terminal !

Thank you sooo much for making this happen! very happy here

The is the beyond of cool. Is the hardware real?

Yep it a AlphaServer ES40 with 4 cpu's, 2GB ram

https://decuserve.org/online/tidbits

> After VSI hosted EISNER for several years, EISNER has now been relocated to Process Software. The Alpha DS20 hardware has been replaced by an emulated ES40 vtAlpha using a license generously donated by AVT and Vere Technologies. VSI continues to provide the complimentary OpenVMS licenses.

I wonder far away QEMU is to supporting OpenVMS on Alpha...

Uh thanks, they moved in MAR 2024 i wonder why they have not migrated to the V9 x86 version instead.

But that's probably why the Uptime changed, before it was always around 300-500 days, now it's around 5-14 day's ;)

>I wonder far away QEMU is to supporting OpenVMS on Alpha...

It's not great:

https://raymii.org/s/blog/The_sad_state_of_Alpha_emulators_f...

Alas. My account has expired.

that box seems to already have the hug of death

sys$system brings tears to my eyes...

Reply on news.ycombinator.com