So, I won’t be wasting too many words on this – partially because I’m not into cheap soap operas, and partially because there’s no way to know what’s going on with this nonsense without dedicating a year’s worth of detailed study into the subject. So it seems that the company Hyperion, which develops and owns the rights to Amiga OS 4 and Amiga OS 3.2 has gone into bankruptcy proceedings. The main shareholder of Hyperion, someone named Ben Hermans, has apparently set up several shell companies (or something?), and they might now own the rights to the two variants of Amiga OS, or they might not? And those shell companies have also gone into bankruptcy proceedings?
Hyperion has been managed by a receiver since last week (Update)
“Ben Hermans BV” (hereinafter: BHBV) is a private company with limited liability owned by Ben Hermans, which has held 97% of the shares in Hyperion since 2019 and acts as a ‘director’ of Hyperion on paper. In March, bankruptcy proceedings were initiated against BHBV for the second time. In the same month, Ben Hermans had already initiated the founding of a new company with the same name.As BHBV has not published any statutory annual reports since 2021, it is currently unclear whether the company still holds the majority of shares in Hyperion. Ben Hermans has not responded to an inquiry from amiga-news.de; the appointed liquidator Charlotte Piers tells us she’ll get back to us in the next few days with “a more detailed response”.
Amiga-news.de
I stopped trying to keep track of this stuff years and years ago, but bits and bobs I’ve picked up since is that there’s been countless lawsuits flying back and forth, questions of rights ownership, and all sorts of other drama you can only keep track of by following the various different Amiga websites and forums in great detail on a daily basis.
As is Amiga tradition.
Amiga OS 4 is an interesting operating system that I spent some fun time with for an OSAlert review way back in 2009, but at this point, if you’re truly hooked on the Amiga OS way of doing things, just stick to AROS. There’s technically also MorphOS, which is pretty great actually, but unless they sort out their own mess of being stuck to dying PowerPC Macs and move to x86 or ARM, they’re basically on borrowed time, too.
Just as sad as usual in OS4 land.
There are still a few diehards that use it, but those that has classic amigas prefer 3.* in most cases, and those on PPC tend to get less and less new software as times goes on.
The browser situation on os4 is rather dire afaik.
I prefer MorphOS over os4 myself, but each to their own tastes.
AROS on x86 is rather great i think, but you need versy specific hardware to get the most out of it, and i do not know if SMP is even there yet, or protected memory. It might be, since it has been a while since i looked into AROS.
The promised AmigaOS4.2 with SMP seems to maybe never be released. Sad, but what can you do.
AROS is the kind of OS you should run in a VM.
AROS is like early Linux, if you want to run it natively, pick your hardware wisely. Though, in my mind, the best way to run it is on the V4SA via ApolloOS!
Let the Amiga die w dignity. LOL.
This is a case when the brand matter more than its future.
I was under the impression that Cloanto had folded all Amiga related stuff. But I was wrong. Sad. HaikuOS is probably light years ahead by now.
I am sure you know this but HaikuOS is the continuation ( well the continuation of the tradition ) of BeOS. The AmigaOS equivalent is MorphOS.
The whole point of MorphOS is to be as close to AmigaOS as possible, if you want a lookalike of AmigaOS that runs on x86 computers there is AROS.
The problem with both AmigaOS and MorphOS is that they have lots of PPC assembly code that can’t be ported to x86 without a rewrite, and they also target binary compatibility with PPC AmigaOS software. BTW I am surprised Hyperion still thinks a business model can be sustained with these restrictions and not closing up shop and releasing everything as open-source or at least freeware.
BTW both AmigaOS and MorphOS run on the brand-new Sam460 boards from ACube systems, which means you don’t have to go to eBay looking for retro hardware, but still, telling someone to shell nearly 890 euros (which is around 970 USD at current exchange rates) just to run an obscure OS is a tough sell.
The sam460 is also almost 15 years old now. That thing should be like EUR50, not EUR900.
Well, that’s the problem with lack of economies-of-scale, everything is expensive because and any fixed costs must be spread across a small number of units.
But again, that’s the nature of running real Amiga OSes (aka AmigaOS and MorphOS): these are OSes with lots of PPC assembly code and as a result require PPC hardware. Long-term we may get cycle-accurate PPC emulation on x86 that will allow running those OSes on standard PCs (much like PCem can emulate a Pentium II PC with a Voodoo 2 in a cycle-accurate manner if you have a powerful enough PC) but as of today OSes like MorphOS and AmigaOS are a hard sell.
because and any fixed costs = because any fixed costs (sorry)
kurkosdr,
With emulation, the question one needs to ask is: why?
Computing needs essentially a purpose. “Why am I sitting in front of my monitor, and typing these?”
For emulation the answer most of the time is “software preservation”, and sometimes “development for other hardware”. And most of that software that is preserved is either closed source industrial ones, or more likely gaming.
Circling back, the Amiga gaming emulation is pretty much a solved problem at the moment. Even a low powered handheld can do it without issues. And yes being “cycle-accurate” is a bonus, but not a necessity.
That leaves development for a non-native OS targeting an emulator a very fruitless task.
Cycle-accurate is not a bonus, it’s a requirement to have accurate software preservation. For example, several PS1 games that didn’t work right under ePSXe (and I hadn’t tested on real hardware since my PSone is in my parent’s house) worked fine under Beetle PSX. At first, I thought the .bin files were bad! Turns out cycle-accurate emulation matters.
But in order to get cycle-accurate emulation, you need a native OS. It’s what those “BIOS files” emulators like Beetle PSX and PCem require are.
PCem in particular requires both BIOS dumps from motherboards and a native OS (for example MS-DOS or Windows 9x), but once you get it running you have a retro PC on your laptop with no emulation issues whatsoever (assuming your computer is fast enough to not get de-syncs of course, but any modern gaming laptop should be able to emulate at least a Pentium 200 MMX and even a base Pentium II).
kurkosdr,
You make valid points. It’s definitely possible to write software that is extremely susceptible to hardware “phase” differences. Even a slight interrupt timing difference between system times/sound cards/video cards/etc could break something in very fragile/buggy software that makes bad assumptions. Correct software is tolerant of synchronization differences, but I have to concede that a lot of software may not have been correct and was only tested on very specific hardware & OS variations. Even though lots of old software may have these ostrich algorithm bugs, for archival reasons we still want/expect emulators to be able to run the software anyways. This can be very challenging.because cycle accurate emulation means you cannot simply implement different subsystems as discrete modules, they all need to be carefully interlocked, which may make it more complex than the original hardware.
kurkosdr,
I would expect that to be true. But at the same time, the operating word here is several. Most games, productivity software, or commercial applications do not need that level of support.
And if they do, we can focus on them individually. Most likely that region of code might require a special handling by the emulator.
That is why they have these lists:
https://rpcs3.net/compatibility
In the extreme example, the NES/SNES cards had additional hardware in them. Meaning just emulating the code did not cut it, as the code was only part of the equation.
And another advantage of not being exact in emulation is upgrading existing software. For instance, Xbox One/Series consoles will double or sometimes quadriple frame rates, increase resolution, enable more post processing in existing games. Again that means some will not be compatible, while others would be much better than the originals.
(Why do you want to run a game in 480p flaky 30 fps when it can run 4K with stable 120)?
>if you’re truly hooked on the Amiga OS way of doing things, just stick to AROS.
Unfortunately AROS is not even a usable replacement for ancient AmigaOS3.1, let alone for anything from the last 30 years.
>Long-term we may get cycle-accurate PPC emulation on x86 that will allow running those OSes on standard PCs
We have that already, OS4 can be run on WinUAE etc. Probably not cycle-accurate but OS4 and its apps generally don’t require that anyway.
Technically the PPC code should be portable to the Talos Workstation. Then you have a very modern and open system to run on.
So, instead of telling people to buy a Sam460 board as they do now (which is available for EUR890 and even ships with AmigaOS 4.1 bundled), Hyperion and the MorphOS people should tell people to buy an exotic workstation that costs anywhere from 5 to 10 grand instead? How does this solve the price-of-entry issue?
Sure, you can theoretically use the Talos Workstation as a daily driver if you only use FOSS software and don’t mind compiling stuff yourself, but let’s be real, for most people, a regular PC (or Mac) and a Sam460 board is a much better deal in multiple ways. The problem with MorphOS and AmigaOS is not the lack of computers available in new condition, it’s the fact you have to buy a new computer to run them.
kurkosdr,
It’s not exactly the same, but I actually feel the same way about mobiles. I imagine more people would be interested in trying alternative operating systems like e and lineageos on their old unsupported phones that are about to get thrown out. But they cannot since the interoperability of ARM devices is so poor.
Isn’t it funny how computers built around the “evil” x86 ISA have standardized bootloaders and you can get 3D acceleration, sound, and even WiFi working on some of them with open-source drivers, yet computers built around the “open” ARM ISA are a mess of undocumented (and mutually incompatible) bootloaders, and require proprietary drivers (and even forked kernels) to work?
Personally, I am the kind of person who believes the importance of the ISA is overblown (unless we’re talking about truly bletcherous ISAs like VAX or iAPX 432). The fascination with ISAs is something pushed by academics that care about this kind of stuff but is not that important in the real world. What gives a computer its personality is what’s around the CPU (including the app ecosystem), not the CPU itself. In other words, all modern architectures can have good and bad implementations and it eventually boils down to backwards compatibility with a given OS’s apps. Even when it comes x86, I am not entirely convinced the massive ARM performance-per-watt advantage we see today is due to inherent inefficiencies of x86 or Intel being stuck on old process nodes. After all, Intel was reigning supreme in AnTuTu benchmarks in phones not too long ago (not enough to justify the compatibility breakage with Android’s ARM native apps mind you, but it shows what x86 can do on a good process node). The only ISA transition I fell was justified was Apple moving away from PowerPC, since it was apparent by then that the AIM coalition was dead. The move to Apple Silicon is also justified, but only from a business perspective (aka Apple owning the entire stack, otherwise they could have cut a deal with AMD or buy VIA technologies and keep using x86 CPUs built on a modern process).
So, I don’t understand this recent desire to run Desktop Linux or Windows on ARM.
But anyway, back to phones, yes, that’s true, modern phones are meant to run whatever OS their manufacturer has flashed from the factory.
kurkosdr,
Indeed.
I agree, an ISA that is good enough and has an app ecosystem is far more appealing to most consumers than an ISA that is theoretically better yet has no ecosystem.
ISA complexity does contribute to some unnecessary overhead because more clock ticks and transistors are needed to fetch instructions into the uop cache. All else being equal this causes higher fetch delays and energy consumption than a simpler ISA that needs fewer clocks and transistors. But whether this is actually noticeable or not depends on the load of the pre-fetcher relative to other subsystems. With tight code loops the prefetcher might actually stay idle most of the time, in which case its overhead is negligible. On the other hand for tasks where the uop cache needs constant replenishing (ie heavy multitasking), the x86 prefetcher will be under a much higher load and could even bottleneck the rest of the CPU. In short, it’s largely going to depend on the load.
In any case I agree that node density has been a much bigger factor.
Well, ARM devices do typically bring better battery life, consumers often benefit from more competition, not having all our eggs in too few baskets is another. Though I agree with your overall premise that these all take a back seat to being confident that the hardware is going to work before you buy it and not having to stress over software compatibility today or tomorrow.
Alfman,
I think we might have missed the boat on this one.
In the x86 era computers were more of an hobby than a consumer device. Today someone buying an android phone, or even a raspberry pi … latest models might not be too interested in portability or open hardware.
The only solution is making a loud noise about this, hence I repeat the concern whenever there is a chance.
suku,
Yeah, manufacturers have learned that empowering consumers in terms of after market repairs/upgrades/support goes against their interests. This isn’t even unique to mobile phones, all industries have learned that customers are more profitable when they cannot help themselves and it’s become an engineering goal.
It’s very different norms. Hardware used to be designed to be repairable/expandable rather than disposable. I believe there still would be just as consumer interest as ever but the technical difficulties are a huge impediment even for technical people
I have very little faith that companies will do the right thing. Even when there’s an olive branch, we’ve been moving in the wrong direction for so long it’s just not enough.
All you achieve with this is preaching to the choir. The Samsungs and Motorolas and Googles consider the non-standardization of ARM smartphones to be a feature, not a bug. It allows them to lock hardware to a given OS (Android variant). Even the Raspberry Pi company enjoys the fact their ARM boards are the ones with the most OS images and community support. If you’ve ever used ODROID ARM boards to run Desktop Linux, you know what I mean.
Alfman,
Yes, even the tractor manufacturers are infamous for locking people in to their ecosystems.
kurkosdr,
I try to bring this up in many contexts. But yes, our voices echo in our chamber.
There is stil “AROS”, which is closest to an open source Amiga OS.
They are still active, but barely:
https://github.com/aros-development-team/AROS/graphs/commit-activity
And even have a Raspberry PI native version:
http://www.aros.org/el/nightly1.php
The issue is, like many other alternative operating systems, there does not seem to be enough community or customer interest to get them going. Still I will try to at least boot it on my RPi to give it a chance.
As a old Amiga user who does have a old Mac Mini with MorphOS to play with, the only thing I find of interest that embodies the best Amiga spirit today is the Vampire: https://www.apollo-computer.com/
It is trying to be 100% compatible with OS 3.x in hardware, has a open source ApolloOS (based off AROS), and is expanding the 68k lineage with the 68080 and various new chipsets (such as Maggie for 3D). It feels like it is taking the next steps the Amiga would have if it continued down the 68k path, where most of Amiga’s true history existed. So to me that is the most Amiga like in spirit because it is being done in hardware where a lot of the Amiga magic was with its custom chips, and then using the only mature open source Amiga like OS available as the OS base from which to expand on.
If you want a stuck in time OS 3.x experience that is mostly compatible then UAE is a good option. If you want to play with one of the better PPC Amiga like OS which actually has a semi-modern web browser (Wayfarer) and looks nice plus is available on cheaper more available old Apple PPC hardware, then MorphOS is a good choice. The only reason I even tried MorphOS was because it ran on old Apple hardware, I would have never bought expensive custom PPC hardware just to run MorphOS. But if you want something that is truly trying to move things forward in Amiga like spirit through custom hardware and OS, then Vampire is it. To me expensive PPC hardware offering nothing more then a PPC CPU over X86 CPU on a standard motherboard doesn’t have the Amiga spirit, and the OS alone (AmigaOS 4.x) is not going to be the the deciding factor because the Amiga was always about custom hardware and software and only Vampire comes close to that.
Already got a MiST but the Vampire looks interesting for running modern stuff.
MorphOS seems to be the only Amiga-like OS with a modern browser (called Wayfarer).
Here is a comparison of AROS, MorphOS and AmigaOS 4: https://eylenburg.github.io/os_comparison.htm