I am an enthusiastic fan of VMS - my first job was tending to a small VAX cluster. I sat down and read the entire hardcopy documentation, which was maybe as many as twenty 3" binders.
OpenVMS got back to me, asking me to clarify just why I wanted this software. I has already explained my background and intent. They said it was insufficient justification and pointed me to a (windows based) student package :( Very discouraging.
I just can't understand why they would make this ancient OS so inaccessible. It is genuinely great stuff, and with an active community it would probably re-emerge as a legitimate contender.
So, all that being said, I followed the link and read the FAQ, and found this:
>"You can participate in the field test if you are a customer or partner of VSI with a valid support contract."
I just can't understand why they would make this ancient OS so inaccessible.
The OS is used by a limited and probably dwindling number of vendors of specialized, complex and long-lived software originally developed for VAX/VMS. I was tangentially related to users of such software for a while; the vendor made the VAX->Alpha and Alpha->Itanium transitions, but saw the writing on the wall and ported the software away from VMS. VSI is not a large outfit, and they're a business. They don't need the hassle of dealing with hobbyists.
It is genuinely great stuff, and with an active community it would probably re-emerge as a legitimate contender.
It's a fascinating and superbly documented system, agreed. But a contender? Far too late for that, IMO, if it was ever in the cards (I don't think so.)
Maybe security through obscurity? They can legitimately tell their customers "not just anyone can get a copy of VMS" and then they can imply that that means that hackers can't hack it or something. I'm sure big government customers go for that sort of thing.
Older versions of VMS have a showstopper login bug that can be used to obtain the full privilege of the kernel.
There are many older VMS installations where this cannot be fixed, and more still where it is not worth the trouble of engaging a support contract to obtain the patches.
Anyone can get a copy of VMS, just not this closed beta.
VSI plans to add the x86 version to the free community license program [1] at a later date.
In the mean time, the Alpha version runs very well under emulation, and is available both under the CLP (free, registration required[1]; also includes Itanium, though AFAIK no emulators capable of running it exist), and as a freely downloadable (no registration required), preinstalled student kit bundled with a preconfigured Windows emulator[2].
Note that the "student" license is unrestricted for non-commercial use; you don't have to be an actual student to use it.
The documentation was amazing. I did campus backups in the white room back in college, and for some reason that was the place they had shelf after shelf of these binders; while the Vaxen hummed, I read through them. It was something of a letdown to move to MAN pages when we got some UNIX boxen.
ah, a fellow tape monkey. Our "white room" had collected 25 years of old hardware and manuals back in the 90s. I loved trawling through it all. Best job ever.
I had no problem getting both Alpha and Itanium kits from VSI, after explaining that I planned to run on an Alpha emulator and a hypothetical Itanium box that I didn't actually own yet, but may someday acquire. So they're clearly not too picky about hardware ownership.
They did initially suggest the student package to me as well, but after replying that I had prior VMS experience and would prefer to install and configure the system myself, they promptly issued my licenses.
Thanks, I gave up after the student package referral... which I guess worked as designed! Maybe now I will get some genuine hardware and take another shot.
Last I heard, VSI's agreement with HP only permits them to issue licenses for VSI OpenVMS releases, and they've never released a VAX build. So at this point, any VAX license would have to come from HP (who, AFAIK, is no longer issuing them, at least not to hobbyists).
I've actually got a VAXstation 3100 M76. It's fun to run it on the hardware, but most of these little VAXen have performance most charitably described as pedestrian.
Yeah they later ones aren’t priced to bad on eBay. You can occasionally find some minicomputers and some of the larger VAXStations but they tend to be a lot more especially considering it would have to ship freight.
I was tempted (but fortunately didn’t succumb) to buy the IBM 4000-series minicomputer my college was decommissioning in the 90s. It would have been both cool and an overwhelming burden to have this thing.
Older versions of z/OS and OS/390 are easy to come by online [0], and from what I've heard run well under Hercules–although I myself have never tried (I have run the public domain MVS 3.8J in the past, in particular the TK4- distribution of it.) Of course, downloading and running those online versions violates IBM's license agreements and quite likely violates copyright laws – but you can make your own judgement about the legal risk and ethics of that.
I think getting your hands on the most recent versions – such as the latest z/OS 2.5 – is a lot harder, unless you work for IBM or an IBM customer, or you know someone who does (and who is willing to help you out, an action which may involve some risk for them). In recent years IBM has added hurdles to make unauthorised distribution harder (such as encrypting the ADCD distributions), although those hurdles are not insurmountable, they just add more work (usual story with DRM). And I think you'll find Hercules will barf on the latest versions of z/OS anyway.
If you want to latest versions from IBM totally legitimately – IBM will sell you a development environment (zPDT) for US$5,540 a year. If you need it for work, could be a totally reasonable investment. If you are just doing it to muck around – well, that's an expensive hobby, although certainly there are people out there who spend more than US$5,000 a year on their hobby, and many people here earn enough they can afford it – but you'd have to really enjoy it for it to be worth it. I doubt there are many people out there who enjoy z/OS that much – although there is one retired IBM guy in Orkney, Scotland (Colin Paice), who does, and he's actually blogged about how he forked the money out and bought it [1].
[0] I think it is better not to say where publicly, in part because doing so draws attention to those places and attention might result in IBM issuing takedown notices. If you know the right search terms it is easy to find.
^ This. The amount of hoops for an aspiring mainframer that just wants to poke at z/PDT and throw money at IBM is silly. They ghosted me after trying to sign up for it.
All you have to do is go to this website – https://www.ibm.com/marketplace/purchase/configuration/en/us... – sign in with your IBM ID (anyone can create one), and provide your details, and give your credit card number. I've successfully done it up to the step where you give your credit card, I haven't taken it further because I'm not sufficiently interested in this to justify to myself giving IBM so much money.
One thing I noticed – IBM does not want individual consumers as customers, only enterprises. If you say they live in the US, the order form says "IBM products are only available for commercial, academic, public, research, or other organizational use", and asks you for an organisation name. I am in Australia, so I give my country of Australia and home address, it then asks me to certify I am buying it for a business, and won't let the order process continue unless I tick it. For Australian orders, it also asks for an Australian Business Number (ABN), which you get from the government when you register a business with them. (Government agencies, and educational/religious/charitable/nonprofit/etc institutions, and trusts, are also eligible for them.) I guess the US equivalent would be an IRS EIN, or a DUNS number, but the US order form doesn't ask for one, just an organisation name.
Anyone in Australia, can get an ABN, but you have to tell the government you are running a business, and in theory they can revoke it if you aren't a real business (and occasionally they actually have). However, another way to get an ABN – anyone can incorporate a company, and when you do you get an Australian Company Number (ACN), and anyone with an ACN is automatically entitled to an ABN (the ABN is just the ACN with a couple of checksum digits prefixed). I actually own a company, which does nothing and has never done anything – one day I was daydreaming about starting my own business, so I created a corporation, and then I daydreamed a bit more, and then I parked the idea – but I pay US$200 a year to the government to keep my corporation alive, so I have a legitimate ABN to give them through that. There is no legal requirement for a corporation to do anything, a "shelf corporation" which stays forever dormant and just pays the annual fees (and engages in a few other pro forma annual legal rituals) is legally just as valid as a billion dollar company
If I gave them my credit card number, and let them charge all that $$$ to it – I expect they'd process the order. How are they supposed to know that my "business", despite being a real legal corporation, is actually just a "fantasy daydream business" not a real one? I don't see how they could tell. (I mean, I could be an independent contractor/consultant/freelancer working through that corporation, rather than a full-time employee – and maybe one day I shall be.) They could cancel the order and give me the $$$ back instead of fulfilling it, but I'd be surprised if they did that.
Indeed, ex-IBM employee Colin Paice in Scotland reports successfully buying it from them – https://colinpaice.blog/2020/06/29/getting-z-os-installed-on... – I guess one thing that may help – don't say it is for a hobby or personal use, claim it is for a business (which if you think about it, doesn't even have to be a lie – tell yourself a story about how you are going to use this to develop software to sell people, and if you believe your own story, that makes it true – who knows, believing in it might even make it happen – and if it doesn't ever actually happen, well a lot of businesses fail, often even just as they are getting started...)
Yeh, the last OpenVMS release for VAX was 7.sommat, and hasn't been updated since. Alpha and Itanium still have active releases with mostly feature parity, tho.
Now if we can only get HP to reinstate the VAX hobbyist license... The VMS Software team can't do that, they never got the rights to the VAX implementation.
Why would they bother? It's a dead platform and hobbyists aren't going to pay big bucks. All the VMS licenses you need are a couple of google searches away, anyway. It was "liberated" years ago.
I think the logic applies for as long as they have customers using the platform. If there are paying customers on ancient hardware, or emulators thereof, then I would expect it to be in their interests to have a way to get new blood to support that.
There was a large VMS software shop in the area. They had moved from VAX to Alpha in the mid 90's. I remember some dude there, back in the mid 90's, showing me a 4 CPU system with a gigabyte of RAM. That was unheard of then. DEC did some amazing stuff. When Compaq acquired it, it was a sad day indeed. They made PC clones, while DEC had their own CPUs and operating systems.
After the Alpha was canceled, the local shop gave up on VMS, ported everything to Linux and IBM AIX, and moved on. Seems like a wise move.
VMS has been dying for decades. I suspect any serious business has ported their app to another platform.
I haven't done it myself, but from what I've heard it isn't hard to work around the licensing restrictions and run it anyway even without a hobbyist license.
I know a lot of people in the OpenVMS hobbyist community were dead set against people doing that, and I do understand why – one of the big reasons was they were worried that it might give the hobbyist community a bad name inside DEC/Compaq/HP(E), which in turn might lead to the hobbyist program being discontinued.
But, when it comes to OpenVMS on VAX – there is no alternative, I really doubt anyone who currently works for HPE cares one iota, and I'm not sure why VSI should be upset about it either? So maybe that former logic is no longer applicable.
Different story for Alpha and Itanium (and soon to be x86), for which there is a hobbyist program run by VSI, and the same concern (that hobbyists hacking license restrictions might upset the vendor to the point that they cancel the hobbyist program) applies to VSI as well.
(I suppose one possible reason VSI might get upset with breaking licensing restrictions on VAX, is that any technical workaround for VAX licensing might also work on newer platforms – but I don't know if that is actually true, and even if it were, VSI could easily make it false by moving to an incompatible licensing system for their OpenVMS releases.)
Why VAX over Alpha (a pure 64-bit RISC ISA)? Do you specifically have VAX boxes to use this on? Regardless, I wish the Deathrow Cluster project was still around (thanks to Da Beave for running it for years).
DEC would have been profoundly wise if they had recognized and adopted ARM, porting VMS to it.
Maybe Cutler would have stayed in that early fait-accomplis. They also never would have spent the money on Alpha, or wasted time selling MIPS workstations.
One thing is that there is a much larger array of interesting old software for the VAX rather than Alpha. The same will be true for Alpa compared with x86, but at least it will be easier to run more modern software on x86.
(I own both VAX and Alpha hardware and run VMS on both)
"Please note that in accordance with the license agreement between VMS Software Inc. and HPE, VMS Software Inc. are not able to distribute VAX licenses."
Is it because they can't distribute licenses for non-VSI versions of OpenVMS? Depending on their license agreement maybe VSI can do new VAX releases, but since OpenVMS VAX system still kicking are running at best a 20 years old release that's been out of support for a decade, even if that's the case they probably think it's not worth the hassle.
Yes I believe that is basically it, their deal with HPE says they can only issue licenses for OpenVMS on VAX if they create a VSI OpenVMS on VAX release, but it is commercially unviable for them to do so – there is no way there are a sufficient number of customers willing to pay a sufficient amount of $$$ to make it worthwhile. Very few people use OpenVMS on VAX in production any more. I think, if it survives anywhere, it would likely be in embedded / industrial control applications – some MicroVAX in a factory somewhere running an ancient VMS version, controlling some machinery – but why would someone want to pay to upgrade such a system to a newer VMS version? They'd get nothing out of it and it would risk breaking things.
Putting that aside, I think a VAX release is also technically much more challenging than Alpha/Itanium was. From what I've heard, the VAX release used different (older) build tools, etc, and even if HPE or VSI still has the source, it would be quite an effort to work out even how to compile it all. By contrast, HPE was still building new Alpha/Itanium versions internally at the time it started to transition to VSI, so it was just transitioning a currently active process to another organisation, as opposed to resurrecting something which hadn't been actively done in many years.
That used to be a very popular naming convention in the 80s and 90s, when "open standards"/"open systems" was a big craze. The term "open source" was only coined in 1998 (or maybe late 1997?), and much of the term's adoption didn't happen until the first decade of this century.
Proprietary operating systems which added POSIX/Unix compatibility frequently used "Open" as branding, either for that compatibility support or for the OS overall. As well as DEC/Compaq/HP(E)/VSI OpenVMS, there is also OpenEdition, which is the IBM's original name for the Unix compatibility component of their MVS mainframe operating system–MVS was later renamed OS/390 and now z/OS (although to be pedantic about it, OS/390 and z/OS refer to the operating system as a whole, officially MVS is just certain of its core components, and as such MVS still exists today as part of z/OS), while OpenEdition was renamed to Unix System Services (sometimes abbreviated to USS, even by IBM, although purists object to that, since USS was already in use as the abbreviation for the Unformatted System Services component of IBM's VTAM mainframe networking package.) Due to this Unix compatibility component, z/OS is actually a certified Unix (UNIX 95). It is also somewhat more than just a compatibility component, in that many newer OS services require it to run – basically anything written in Java, or anything which is shared code with non-mainframe IBM platforms, runs under it. A lot of the systems management frameworks in newer z/OS releases can't run without it.
There is also a POSIX compatibility component of IBM's VM/CMS operating system (now known as z/VM) - originally called OpenEdition too, later renamed to OpenExtensions. OpenEdition/VM was always significantly more limited than OpenEdition/MVS, in part because Unix APIs were technically more challenging to implement under VM/CMS than under MVS (since native VM/CMS is further away from Unix than native MVS is), in part because less was invested due to less market demand at the time (and very little remaining today). OpenExtensions is basically in deep maintenance mode, IBM still supports it but have publicly said they have no plans to add new features; OpenEdition is in somewhat of a shallower maintenance mode – they aren't interested in updating it from UNIX 95 to a newer UNIX standards release (UNIX 98 or UNIX 03 or UNIX V7), but given they actively use it to provide higher level OS functions (such as Java support), inevitably some enhancements sometimes get made to support new requirements in those areas. Generally, for anything more demanding, IBM will recommend using z/Linux (Linux on the mainframe), or zCX (a port of Docker to z/OS, wherein z/Linux is run under a hypervisor which is built in to the z/OS operating system.)
Similarly, when Siemens added POSIX compatibility to their BS2000 mainframe operating system (mainly used in Germany), they renamed it to BS2000/OSD (OSD = Open Server Dimension). BS2000 is still around, although Siemens ended up selling it to Fujitsu.
Fujitsu owns at least three different mainframe lines – as well as BS2000, they also bought the British firm ICL who owned the VME mainframe platform, and they have also have a "clone" of IBM MVS, MSP. The "clone" is not really a clean clone, they actually stole a lot of IBM's code to build it. IBM sued them (and Hitachi too who were doing the same thing), and the lawsuit settlement included a license from IBM for them to legally continue doing it (for which Fujitsu and Hitachi paid IBM several hundred million dollars each). MSP is less than 100% compatible with MVS, and only with old 1980s/1990s vintage 31-bit MVS. I believe it currently only survives in Japan, although it used to be popular in some other countries too, in particular Australia. A lot of Amdahl machines ran Fujitsu MSP, and then Fujitsu bought Amdahl. But, back to the topic of Open – when ICL added a POSIX compatibility subsystem to VME, they renamed it to OpenVME. And OpenVME is still around too; its major remaining use is inside the UK government, which has various core systems trapped on it.
After learning about the OpenVMS community license I got excited and applied for a license. Bought myself a [PiDP-11](https://obsolescence.wixsite.com/obsolescence/pidp-11) and got ready to party.
OpenVMS got back to me, asking me to clarify just why I wanted this software. I has already explained my background and intent. They said it was insufficient justification and pointed me to a (windows based) student package :( Very discouraging.
I just can't understand why they would make this ancient OS so inaccessible. It is genuinely great stuff, and with an active community it would probably re-emerge as a legitimate contender.
So, all that being said, I followed the link and read the FAQ, and found this: >"You can participate in the field test if you are a customer or partner of VSI with a valid support contract."
What gives, VSI. what the heck even gives?!