Back when I was working on my article about PA-RISC, HP-UX, and UNIX workstations in general, I made extensive use of OpenPA, Paul Weissmann’s invaluable and incredibly detailed resource about HP’s workstation efforts, HP-UX, and tons of related projects and products. Weissmann’s been doing some serious digging, and has unearthed details about a number of essentially forgotten operating system efforts.
First, it turns out HP was porting Windows NT to PA-RISC in the early ’90s.
Several magazine sources and USEnet posts around 1993 point to HP pursuing a PA-RISC port to NT, modified the PA-RISC architecture for bi-endianess and even conducted a
↫ Paul Weissmann at OpenPAback-roompresention at the ’94 Comdex conference of a (modified HP 712?) PA-7100LC workstation running Windows NT. Mentions of NT on PA-RISC continued in 1994 with some customer interest but ended around 1995.
The port eventually fizzled out due to a lack of interest from both customers and application developers, and HP realised its time was better spent on the future of x86, Intel’s Itanium, instead. HP also planned to work together with Novell to port NetWare to PA-RISC, but the work took longer than expected and it, too, was cancelled.
The most recent secretive effort was the port of HP-UX to x86, an endeavour that took place during the final days of the UNIX workstation market.
Parts of the conversation in these documents mention a successful boot of HP-UX on x86 in December of 2009, with porting efforts projected to cost
↫ Paul Weissmann at OpenPA100M+between 2010 and 2016. The plan was for mission-critical x86 systems (ProLiant DL980 and Superdome with x86) and first releases projected in 2011 (developer) and 2012 (Superdome and Linux ABI).
I’m especially curious about that last one, as porting HP-UX to x86 seems like a massive effort during a time where it was already obvious Linux had completely obliterated the traditional UNIX market. It really feels like the last death saving throws of a platform everybody already knew wasn’t going to make it.
This is a common theme for aging incumbents in an established industry when a new upstart technology comes to make all of them obsolete.
The most famous example is Kodak inventing the “Charge Coupled Device” (CCD, the original digital camera sensor). But shelving it to focus on their thriving film business. As they sat on the patent to expire, others ate their lunch, and film is but a niche product now.
Same with “dinosaur” UNIX vendors. Even when Linux was ramping up, they had the change to move their systems to open source, and free themselves from software maintenance burden. Some tried to open source their own offerings (Solaris), a bit too late. Others actually adopted Linux to their ecosystem (IBM/AIX), but most of them just “honorably died fighting”.
(There is also the issue with funding the team which no longer is relevant. When traditional auto makers, for example, tried to switch to EV they failed, except maybe volkswagen which did not have a choice. As they had to drop their ICE teams, their part vendors, and entire dealership networks to “start fresh”, which would receive massive backlash from VPs and teams that are about to lose their jobs).
I am not sure whether open-source plays a role, after all MacOS X isn’t open source (not fully) and is doing fine. Also, Oracle closed Solaris after realising that open-source wasn’t offering them anything in terms of less software maintenance burden because the open-source community was pre-occupied with Linux. But one thing that was certainly the kiss of death for those Unix OSes was the reliance on ISAs whose implementations offered markedly less performance per buck compared to x86 offerings (despite being RISC CPUs and theoretically capable of offering more). At the final years of the Unix workstation, you bought a Unix workstation because you really wanted a workstation with more than 4GB of RAM, it’s the reason the high-end SGI Tezro outlived everything else in SGI’s lineup for example. But the arrival of AMD64 eventually ended that niche too.
Steve Jobs literally saved MacOS X the last minute by going to intel (there was no chance those PowerBook G4’s could compete with Core Duo laptops) and Jonathan Schwartz would have saved Solaris if he hadn’t foolishly spent 1bn dollars acquiring MySQL (which was already open-source anyway) just before the Great Recession hit.
kurkosdr,
Yes, it might not be 100% from open source, but I could argue Linux being available was a large catalyst to downfall of these operating systems.
And two minor corrections:
1. macOS is (was?) open source. At least the kernel:
https://github.com/apple/darwin-xnu
This is how “Hackintosh” developers were able to keep up so far.
2. Solaris tried to move to AMD64 too, but that might not have been the right move either. I remember playing (more like touching) one in our IT department back in the day. The issue was, they also tried to materialize Java on the desktop (which was very slow), and… Intel server hardware was already coming up to be much better than AMD offerings.
Anyway, you are right on MySQL point. It was not helpful. And that meant direct competition with Oracle, which might also had some influence on the buyout.
sukru,
I agree, free clones were definitely key players in the downfall of the commercial unix market. Without free clones, the commercial user-base would not have fled. Ironically good interoperability between unix operating systems made quick and easy migration far more practical. It clearly hurt commercial vendors when much cheaper clones popped up. I think windows could have faced the same fate had there been highly effective clones of it. The closest thing, reactos, never got past alpha quality status.
I don’t know if it is an intentional act of self preservation by Linux leaders or not, but now as the undisputed leader in FOSS, they aim for Linux-specific interfaces rather than portable ones thereby forcing everyone in FOSS circles to copy linux or else become irrelevant. This keeps linux at the front of the pack at the expense of innovation from myriads of other operating systems that don’t go anywhere without linux adoption.
Exceptions like ZFS come to mind, but that’s not technically mainline linux and it was very popular among enterprise users before a linux specific alternative btrfs was developed. So in this sense it’s kind of another example of unix technology that was “not invented here” getting reinvented into a new linux first technology.
It’s hard for me to label this as entirely “good” or “bad”. After all, competing technology is useful to push innovation, but on the other hand, there’s a ton of redundant effort and linux seems to be avoiding synergies that would benefit FOSS as a whole.
Nothing ironic here, Desktop Linux aka GNU/Linux was specifically envisioned as a Unix-compatible OS to ease conversion from Unix, a decision which has its roots in the GNU project being envisioned as a Unix-compatible OS to ease conversion from Unix. In other words, hurting commercial Unix vendors by offering a free alternative to their products was the plan all along.
Yes, too bad the “Thorium core” Kickstarter for ReactOS failed. It might, just might, have gotten somewhere.
Can’t agree with that, Desktop Linux has to catch up with MacOS X, Windows, and even Android to stay relevant. Let’s be honest, no modern OS can compete by relying on init.d (aka a bunch of hacks on top of the shell that uses PID files to track processes) and X.org (just no). If we listened to the FreeBSD people, even D-bus wouldn’t have happened because it’s Desktop Linux-specific after all.
kurkosdr,
You are speaking from the perspective of the FOSS alternatives, but it is ironic that the commercial unix vendors themselves promoted and embraced the interoperable unix standards that were part of their undoing. Say there had been more compatibility issues impeding FOSS adoption…commercial unix vendors would have had more exclusivity to leverage.
DOS clones were also very close on the heals of MS DOS and MS didn’t have much time before free DOS clones would make a big dent in microsoft’s OS marketshare. However pivoting to windows likely saved microsoft’s hide. Windows and it’s caveats were notoriously hard for 3rd parties to replicate precisely. 3rd party compatibility with windows applications is still hit and miss even decades later.
You’re counter example isn’t a FOSS project though. My claim was about linux specifically leading FOSS.
I honestly don’t see how any of that contradicts what I said. I believe *nix should keep evolving. But I do believe we’re missing out on opportunities to combine our collective community resources to evolve while getting more bang for the buck. For better or worse though the linux culture of not-invented-here is a barrier to more cooperative efforts.
Sure, but as Sun found out, open-sourcing your OS won’t make all those open-source developers working on Linux (some of them sponsored by competitors such as IBM) contribute to your Unix OS. Then there is the issue that open-source projects descended from proprietary products usually can’t have everything open, which led to the OpenSolaris/Solaris duality.
Really, the only viable path to profitability for proprietary Unix vendors was offering features Desktop Linux didn’t. Only MacOS X achieved that (by offering a UI, graphics stack, and app ecosystem that Desktop Linux didn’t), and only on the desktop (Apple killed Xserve eventually, and even MacOS X Server edition).
That was what the “not fully” clarification in my previous comment alluded to. MacOS X had a GUI-less and graphics-less “core” server version called Darwin, but Apple no longer provides binaries or build instructions for it and the latest source code dumps are not buildable, so no runnable Darwin OS is buildable anymore. But hey, MacOS X is still partially open-source in some sense.
The move to x86-64 is the reason Solaris is still a thing, with Sparc hardware being dead for at least the past 6 years. The Sparc version was offered alongside. I don’t understand how Java supposedly fits into the whole thing, it was a long-running policy of Sun (way before the decision to port Solaris to x86) to position Java as an OS-neutral platform that runs on Windows, Linux, Solaris, other Unixes etc
“core” server version = “core” version (you can use Darwin as a GUi-less desktop OS, I guess)
The “Java Desktop” was two things.
First a GNOME variant that was heavily modified:
https://en.wikipedia.org/wiki/Java_Desktop_System
Second a completely new system, actually written in Java:
https://en.wikipedia.org/wiki/Project_Looking_Glass
At the time Sun was vehement on getting Java on the desktop, everywhere. That was probably around the time Microsoft “embraced and extended” Java (the infamous Visual J++. I could still argue it was a really good idea, sabotaged by the stubbornness of the both companies)
>”I am not sure whether open-source plays a role, after all MacOS X isn’t open source (not fully) and is doing fine.”
Not in the server space. MacOS X is only doing fine in the retail consumer industry segments.
Using Mac OSX is not really an argument to the success (or failure) of UNIX vendors. Mac OSX was built to replace the outdated and definite non-UNIX “Classic” MacOS, whilst retaining source (for recompilation) and binary (via a compatability layer) compatability with Classic MacOS.
The fact that OSX had a UNIX core was more an added bonus for some. MacOS 9 could have easily been succeeded by a non-UNIX, with BeOS famously being in the running as a successor.
Mac OSX is successful because it runs Mac OSX software, not because it runs UNIX software.
The123king,
I agree it could have been succeeded by non-unix as far as mac software cared. However I actually do think OSX benefited from having unix roots. Not only did this make OSX compatible with tons of unix languages & daemons that mac shops could only use natively thanks to unix compatibility, but I’ve seen first hand just how popular macs were with unix/linux developers. At local linux user groups, scanning across the audience and you could see apple macs were by far the most popular personal computers there. Many technical elites were buying macs to work on unix software, but I don’t know if they would have been as interested in macosx if not for its unix backend.
Alfman,
At one point even Linus was famously using mac as his development platform. (Not sure whether this is still the case though).
I love Open Source but I do not think this had much to do with it. One of the earliest and most aggressive “Open Source” tech companies was Sun. At the time, their rapid move to Open Source was widely blamed for their downfall. Sun revenue dropped by 90% and many critics blamed it on Sun’s Open Source moves. Of course, the real reason Sun failed is because Intel based systems became “good enough” for way less money. That killed off all the entire UNIX server and workstation world. When AMD added AMD64 ( x86-64 ) to allow access to more than 4 MB of RAM, it was all over.
Linux was a factor in allowing commodity hardware to compete of course. But this is also when Windows suddenly emerged as a serious server option as well so Linux was not necessary per se. Windows grabbed way more of the market than Linux did at first. Neither Windows nor Linux were even close to better on the software side at the time. They just ran on dramatically cheaper hardware.
Later, Linux pulled the server market away from Microsoft but I think maybe Apache had more to do with that even than Linux itself. With the rise of “the cloud” and cloud architecture ( containers ), Linux is actually the superior solution. Open Source is an important part of that story but given the amount of money customers are pouring into Azure and AWS and the amount of proprietary lock-in that those platforms create, it is not that simple. Before the cloud, Red Hat was able to secure A LOT of money for their Open Source solutions while other just as capable ( or even largely identical ) software stacks struggled to keep the lights on. Red Hat offered more than just software and for less money than the real competition.
Microsoft avoided the fate of the UNIX vendors because, unlike them, they were able to evolve fast enough. Microsoft Azure is the cash flywheel of the company now and I am sure there is more Linux running on Azure than there is Windows. They were smart enough not to die defending the Windows hill. Microsoft stole the server market from UNIX but then lost the server market to Linux. But who they were REALLY losing the market to was Amazon. They managed to pivot and, while they lost a lot of growth to Amazon, they did not really lose the market. The ultimate winner in that fight is perhaps not yet known.
I have used Linux daily since about 1992 but I know that is odd. Open Source has not helped Linux very much at all on the Desktop. Windows continues to dominate and macOS is still more popular even though Apple barely tries. You cannot tell me that the Darwin kernel being nominally Open Source has anything whatsoever to do with how popular macOS is. If it is just about Open Source, why has Linux not taken off more?
What Open Source mostly seems to do is to BLOCK proprietary software from dominating a new market. Open Source matters most when markets change. Existing Open Source options get taken up by new entrants and the beachhead that creates makes it hard for proprietary offerings to take hold. We see that in mobile devices for example where Android has sucked away the oxygen that competitors ( even Microsoft ) needed. I think that is what happened in “the cloud” as well, especially with AWS which made it hard for even Microsoft to push Windows there despite their dominance in client / server. It is what happened with web servers where Apache made it hard for Microsoft to leverage their server dominance into complete dominance of the Internet with IIS. The existence of Linux has made multiple waves of embedded ( sorry IoT ) a harder and harder place to profit from proprietary software as well. Open Source is rocket fuel for commoditization.
People will use free or even just inexpensive proprietary software without much care. They question is just how the people writing that software make money. Apple gives their software away. It is not Libre but it is gratis. That is because they still make money selling hardware. They have pivoted on the hardware side several times to keep themselves from getting swept away. More and more though, they make money selling services. The “dinosaur” UNIX vendors never found a way to make enough money without selling expensive hardware. Perhaps their shareholders just did not give them enough time.
Open Source did not and would not have saved any of the big UNIX hardware vendors. If Sun could have made cheap computers that ran Solaris, Solaris would probably still be popular. At least, it would likely have lived long enough to fuel or compete with AWS. Nobody ever wants to give up their profits when their markets get disrupted. They try to hang on to margins that cannot be justified and the “good enough” solution eats their lunch. That is how Microsoft succeeded in pretty much every market. They understand that playbook well enough that they knew they could not do that with WIndows. They pretty much gave it away for a while to keep it relevant ( and to keep the developer attention it offered ). Then they pivoted to create Azure where they make money from the same customers in a different way. They can make lots of money with Azure. If what they were trying to do instead was to charge high license fees for Windows Server either on real hardware or AWS, they would be dying.
Porting HP-UX kinda sorta made some slight sense, since Itanium was going nowhere, and HP really needed a replacement architecture for their high end integrity systems, that provide them with a nice revenue stream (specially in terms of support contracts, etc).
Also organizations of the size of HP tend to have massive intertia. So there is a lot of teams within them with an interest in continue their existence. So the HP-UX/Integrity teams most definitively wanted to continue existing, so a port to X86 seemed like a good bet.
While 2009 was way way way too late, not when the project began and IMHO, probably on the edge of the last bit of old Unix bases. I mean, there’s people today, that run those old Sys V.2 variants. Believe it or not.
Pretty typical of closed IP companies. Lacking vision to the point of causing their own destruction (at least they stood firm!!).
Having work with and talked with HP’s “open source” moguls, their time from start to finish was immense (to the point of obsolescence). There’s just no way they could compete, at least with their strange view of FOSS. While they weren’t an enemy of FOSS like IBM, their attempt to “doing it” was never going to work.
Using one’s knowledge of how to build a really nice truck doesn’t translate well into building spaceships.
How was IBM an enemy of FOSS? They were one of the big orgs to bet big on linux.
Organisations change. IBM definitely was bitter to Open Source in the early days. Adopting Linux was more a “sink or swim” choice.
The123king,
That’s a good point. Organizations can change in terms of both strategy and management. Heck even different divisions could be at odds with each other. Long term IBM knows that turning down money is bad business.
IBM has embraced linux. I don’t know that IBM has embraced “open source” though. This is kind of an odd statement to say because linux and open source go hand in hand…. but at the same time I can’t ignore IBM’s influence on redhat, killing the original centos, making FOSS code inaccessible to the FOSS community, and treating RHEL more like a proprietary product pointing everyone to rely on other upstream staging distros for source instead.
I think that IBM has embraced Linux as a commodity. They are comfortable building business on open standards. That is how they think of Open Source.
Open Source is not a strategy for them. I would not expect a significant Open Source product to come out of IBM. Contrast that to Microsoft who, while certainly still highly capitalist, I think understands Open Source these days.
IBM both does way more Open Source than you might think and at the same time nothing of consequence:
https://github.com/IBM
In their own words, “IBM developers produced this code as an open source project (not as an IBM product), and IBM makes no assertions as to the level of quality nor security, and will not be maintaining this code going forward.”
tanishaj,
You may be right, but there’s still a significant difference between the two.
I was disappointed by the lack of source code for ms edge browser. I suspect they wrote a code remover to strip all the code that they were not obligated to provide source for, leaving the edge source project as an empty husk.
Thanks for linking it.