@NutsAboutAmiga

I am quite optimistic about what we now call the amiga. There are going to be a few hiccups along the way, but the future is okay. It probably won't take over the world but will be soon be in its rightful place. There is not going be one amiga but several--each with their own individual characteristics, but with enough commonality to identify them as amiga.

About some individual types now in current vogue--many will disappear. The problem with PPC Amigas is that they are essentially "wrappers" for an emulator. Usually, they are quite fast emulators, but nonetheless they are emulators. I have high hopes for V54,but if the system does not attract competent coders, it will be another failure to produce an advanced amiga.

On the classic side the THE A500 Mini has a lot of appeal. I won't be buying one. I have a queasy feeling about this computer. While there are quite a few hacks for the machine available there are things that don't make me every comfortable. I understand the appeal of the machine . I own a C64 Mini. If you understand the limitations of these types of things and are willing to live with the limitations that that imposes, go ahead and buy one.

Was going to say some things about various options in the way of hardware, but have decided to leave that for another thread at anther time.

@Thread

the last Amigas where build under Amiga Technology until 1997.

Since then no Amigas where build and we should be glad that not some snotty guy came up selling a low quality tablet with UAE under that name.

In my view, the only hope for the platform is to unshackle itself from the burden of obsolete and/or overpriced hardware. You have a difficult intersection of requirements left over from this:

The desire many users have for it to be a "real" computer and not just an emulator running on some other host.

The desire many users have for it to run some "legitimate" continuation of the original operating system.

The desire many users have for it to have backwards compatibility with software they already have.

The desire many users have for it to be performant enough to take on more modern computing tasks.

The desire almost all users have for it to be affordable.

There are other wishes too but these seem to be most common in my experience.

Every offshoot seems to satisfy a few of these desires at once, but usually miss one that's a deal breaker for someone. There was a solution that ticked all those boxes but it was strangled and discontinued.

_________________
Doing stupid things for fun...

@Neuf

Quote:

About some individual types now in current vogue--many will disappear. The problem with PPC Amigas is that they are essentially "wrappers" for an emulator. Usually, they are quite fast emulators, but nonetheless they are emulators. I have high hopes for V54,but if the system does not attract competent coders, it will be another failure to produce an advanced amiga.

In what way? 68K support? The only thing emulated is 68K programs. Everything else in the OS is running PPC native. Which was actually the point of OS4. Had 68K Amiga programs only ran via UAE in OS4 it would have flopped years ago with limited software available and no 68K support.

I don't see V54 really changing anything. But do you mean V54 as in OS4.2 or System54 from AmigaKit. What OS4 needs is built in infrastructure so compiling standards like SSL, or a GUI engine like Qt or any open source code isn't such a roadblock. Along with more affordable hardware.

One thing I suppose I should ask is, what exactly does the poll mean when it says "future"?

AmigaOS as we understand it today has no future to be optimistic or pessimistic about. It is an OS from a different era. A happier time when single user, non memory protected operating systems were commonplace and you didn't have to constantly worry about being the victim of malware thanks to an always connected world. I don't see AmigaOS ever meeting present day notions on OS security, let alone tomorrow's, without changing it into something unrecognisable in the process.

In that regard I don't ever see it as being a mainstream OS competing against the likes of Mac, Windows, Linux etc.

_________________
Doing stupid things for fun...

@Karlos

In the context of the topic question, "Amiga" is a catchall term. It includes all of the past, present, and planned future developments in hardware and software which are related to:
- Classic Amiga hardware running Amiga OS 3.x
- PPC-based hardware running AmigaOS 4.x, MorphOS
- The hinted System 54 for PPC-based AmigaOne hardware
- The will-they/wont-they A1222 Tabor
- Some crowdfunded PPC-based laptop which if it ever comes out might see the possibility of having the chance of maybe being considered for an AmigaOS 4.x port. Or optionally a MorphOS target.
- AROS on x86
- The elusive MorphOS on x86
- AROS on 68k
- FPGA-based 68k recreations such as the Vampire Standalone V4+ with ApolloOS
- Classic Amiga HW rejuvenated with Vampire FPGA-based 68k expansions
- UAE et al
- PiStorm
- The A500 mini
- emu68
- Project deadw00d
- Take 3 on the Amithlon concept

Last edited by agami on 03-May-2022 at 03:55 AM.

_________________
All the way, with 68k

@Karlos

Quote:

Karlos wrote: In my view, the only hope for the platform is to unshackle itself from the burden of obsolete and/or overpriced hardware. You have a difficult intersection of requirements left over from this: The desire many users have for it to be a "real" computer and not just an emulator running on some other host. The desire many users have for it to run some "legitimate" continuation of the original operating system. The desire many users have for it to have backwards compatibility with software they already have. The desire many users have for it to be performant enough to take on more modern computing tasks. The desire almost all users have for it to be affordable. There are other wishes too but these seem to be most common in my experience. Every offshoot seems to satisfy a few of these desires at once, but usually miss one that's a deal breaker for someone. There was a solution that ticked all those boxes but it was strangled and discontinued.

This is a very good summary. I like it a lot.

_________________
All the way, with 68k

@Karlos

I think you've pretty much summed it all up.

Quote:

There was a solution that ticked all those boxes but it was strangled and discontinued.

Which one was that? My mind is going towards AAA and Hombre. Before any post Commodore solutions.

@Hypex

That's an interesting one, for sure. In terms of what I'd written I was reflecting on the post commodore era. You can probably guess I was talking about Amithlon (Umilator to be precise). Admittedly, it's an emulation, but it's not an emulator in the sense of just another application you casually run from Windows/Mac/Linux.

Last edited by Karlos on 03-May-2022 at 05:32 AM.

_________________
Doing stupid things for fun...

@Karlos

"Modernizing" in my sense is what all mainstream platforms for a long time offer and is needed today: 64bit, full memory protection, multicore support, USB3, full 3D and so on

first my general view (as explained already)... There is 68k branch including PiStorm and V4 and there is NG (what is not really NG in todays sense either). 68k is the biggest market with most users, most software and even new development. It is niche but it is living. You have a big software base that (in many cases) is closed so you cannot recompile it. "Modernizing" it would mean sacrifizing lots of software without adding real value to customer. I would keep it as it is-

The wishes might be there but is not realistic in my view except someone with plenty of money steps in and buys out everyone and invests lots of money in development of OS and new software

"The desire many users have for it to run some "legitimate" continuation of the original operating system."
"In my view, the only hope for the platform is to unshackle itself from the burden of obsolete and/or overpriced hardware. You have a difficult intersection of requirements left over from this:"

Both can only be related to AmigaOS 3.X and 4.X... 3.X makes no sense to modernize as written above. 4.X is not legal to be ported to anything else than PPC if the comments from one former amigaos developer is true. And even if that is solved by money it would mean to drastic change to the OS. It would need lots of efforts. And even if the OS would be modern and running on cheaper hardware you still need lots of drivers and (most important) modern software, in best case exclusive software that makes it interesting for people to use it. There is so much missing that I see no chance at all there. It would also mean 32bit 68k amiga software can only be used in emulation.

For me the only idea that makes sense there is using a modern base like linux and add amiga components like desktop and certain software and components. By that you win a modern base with modern software and can get amiga look & feel

BTW the stubborness of many users to stick on the "legitimate" version and hope for a wonder and vague promises instead of using one of the alternatives has led to the current situation.

For me only the 68k branch is promising (as a niche) and Deadwoods project that could led to a kind of linux distribution that includes desktop, components and software from amiga and offers the look & feel and is at the same time useable in todays terms. I will do there a distribution in the future.

About MorphOS X64 I cannot say much. We will see.

AmigaOS 4.X is dead end in my view

Aros is existing on a lot of platforms, 68k is a success, X86 has the same problem as other niche systems (missing drivers and missing modern software). Aros as part of Linux is promising.

Last edited by OlafS25 on 03-May-2022 at 10:50 AM.
Last edited by OlafS25 on 03-May-2022 at 09:13 AM.

@OlafS25

As I mentioned in the other thread I've been posting in, what are the important features of OS4 that are contingent on it running on PPC that couldn't be implemented on 68K? The biggest architectural promises of PPC over 68K are multicore and 64-bit addressing. Neither of these have been leveraged. That leaves performance. No doubt a lot of the UI improvements etc would be a challenge for all but the fastest 68K hardware solutions but certainly not for JIT on x64, especially with the option of native code passthrough for compute intensive code.

I'm sure most of the features that are perceived as important could be reimplemented in 3.2 if porting was a legal impossibility.

_________________
Doing stupid things for fun...

@Karlos

I do not know how easy that would be... Thomas Richter always wrote that the 68k branch of amigaos is not very clean and difficult to improve. But even if... most important in a system used for every day tasks today is security and there full memory protection is a must. That collides with all traditional amiga concepts and would break software

@OlafS25

Perhaps. Nevertheless, OS3.1 has been ported to C once, which became the basis for 4.x. Its been reimplemented from scratch in C twice, for MorphOS and AROS. If there was a will, it could be done. If you want to know who wants the platform to remain dead, it's whoever keeps it tied to a PPC only target.

As for security and other modernisation aspects, I reiterate. The host kernel layer can provide some security but to implement these concepts in AmigaOS itself is the challenge. And it's a challenge NG has not delivered upon either. Doing that would change it so radically I doubt there's much possibility of retaining anything familiar about it. You may as well just use a full Linux with a user interface theme, UAE and maybe some AxRuntime binaries.

Just as a reminder, this notion of a Umilator style solution is just a part of what I see as a continuation of the legacy. The overarching core of it is 68K as a CLR implementation with a set of ABI/API and JIT/AOT support standards, that would allow applications to run on real 68K hardware/OS as well as each NG platform that supports 68K binaries today. Having the ability to run it all on a platform that's not tied to niche or obsoleted, expensive hardware is simply how you'd broaden the user base and mitigate against further obsolescence.

Last edited by Karlos on 03-May-2022 at 12:05 PM.
Last edited by Karlos on 03-May-2022 at 12:01 PM.

_________________
Doing stupid things for fun...

@Karlos

that is what I always say

68k for retro niche, AxRuntime based linux distribution as professional option

your super-amithlon would certainly beat any other 68k platform but how much interest would be there is difficult to judge for me. Current amiga (68k) users primarly want it for playing old games so I am not sure how many would want to use it. And if you really attract new users from outside is at least doubtful to me.

But as I wrote... go for it

@Karlos

Quote:

You can probably guess I was talking about Amithlon (Umilator to be precise).

I probably didn't pick up on that despite recent themes.

I just didn't see it as ticking all boxes.

It had limited backwards compatibility since it was optimised for speed but still running Exec 68K at the core would have increased compatibility with OS friendly software.

Quote:

As I mentioned in the other thread I've been posting in, what are the important features of OS4 that are contingent on it running on PPC that couldn't be implemented on 68K?

The most important feature was the CPU being available. It was the closest to an official 68K replacement and touted as such. Apple chose it it replace 68K when forming the AIM alliance. It was big endian. And it was the only CPU when running inside an Amiga that could clock to more realistic speed of 240Mhz at the time.

Because AmigaOS is a single threaded 32-bit OS. Heck, it doesn't even have threads. It has a shared memory model and they still haven't added threads.

Quote:

The biggest architectural promises of PPC over 68K are multicore and 64-bit addressing. Neither of these have been leveraged.

Last edited by Hypex on 03-May-2022 at 02:01 PM.
Last edited by Hypex on 03-May-2022 at 01:59 PM.

@Hypex

Quote:

Because AmigaOS is a single threaded 32-bit OS. Heck, it doesn't even have threads. It has a shared memory model and they still haven't added threads

Erm, I'm not sure Carl Sassenrath would agree with this. Unless by threads you actually mean separate memory space processes. Which are called processes in POSIX speak. Threads explicitly live in the same memory space as the process that creates them. AmigaOS can be viewed as a single process in a POSIX sense but it definitely has threads. Every exec Task and DOS process that builds on a Task is a thread.

_________________
Doing stupid things for fun...

@OlafS25

Quote:

There is 68k branch including PiStorm and V4 and there is NG (what is not really NG in todays sense either).

That's because NG, in respect to OS4, really goes back to 2004. That's almost 20 years ago. When they got OS4 out for the XE. Back then it was a lot closer to NG as the XE was close to what Apple were producing. Then Apple moved from the G5 to x86. And it left the following AmigaOne models with nothing directly comparable too. With only PC computers on desktops, and obviously improving massively over the years, the NG computers just fell right behind.

Quote:

You have a big software base that (in many cases) is closed so you cannot recompile it.

I think most Amiga software that can be recompiled has been. At least if the author is willing and it is in C or E or some high level language. For ASM, fixing to be OS friendly if it hits hardware, I think is the best you can do and expect.

Quote:

"The desire many users have for it to run some "legitimate" continuation of the original operating system."

I think this matters. Well it matters to me. It's why I invested in the OS4 platform in the first place.

Mind you, there comes a point where you can't leave old code hanging around, and must replace it Similar story, the present code became too hard to maintain, so needs to be completely replaced and rewritten.

At this point, the original authors aren't maintaining the code like they were, so we must expect foreign ideas to be brought in.

Quote:

4.X is not legal to be ported to anything else than PPC if the comments from one former amigaos developer is true.

I only recall Hans-Joerg commenting that they didn't have a license to port to x86 so shut up or something like that.

Quote:

. It would also mean 32bit 68k amiga software can only be used in emulation.

I don't see that as an issue. The fact is on OS4 68k code IS emulated. So all this talk about breaking backwards compatibility just looks like hand waving crap to me.

What they needed to do is create a whole new fresh new API based on the old one that didn't have the legacy problems. Or at least could be used as a bridge for breaking the 64 bit barrier and going multi core. Then just expose a 68K API that internally called the new API cleanly. Of course porting Amiga apps would have been harder than those damned interfaces. But, in the important sections, such as new functions to access system objects and banish Forbid locks forever!

Mac OS software can work on OSX fine, after the quirky booting, and those are two completely different OS. So no reason OS4 can't do it. And yes, that is considering sandboxes and the more direct approach OS4 has for 68k support which I did like, but emulating 68K interrupt code inside a real PowerPC interrupt is a bit far.

Quote:

For me the only idea that makes sense there is using a modern base like linux and add amiga components like desktop and certain software and components. By that you win a modern base with modern software and can get amiga look & feel

So do what Apple did? Just dump the old OS. And create a new one by skinning it on top of another kernel?

Quote:

BTW the stubborness of many users to stick on the "legitimate" version and hope for a wonder and vague promises instead of using one of the alternatives has led to the current situation.

But what is it an alternative to? This is similar to what is a real Amiga and the many different answers. Without any legitimacy, anything else is a fan copy, no matter how good it is.

@Karlos

Quote:

Erm, I'm not sure Carl Sassenrath would agree with this. Unless by threads you actually mean separate memory space processes. Which are called processes in POSIX speak. Threads explicitly live in the same memory space as the process that creates them. AmigaOS can be viewed as a single process in a POSIX sense but it definitely has threads. Every exec Task and DOS process that builds on a Task is a thread.

The whole AmigaOS is a process and programs are the threads joke doesn't help here. AmigaOS doesn't do fork(). And the pthreads variant on OS4 is fake. In the modern sense sense you need copy on write mechanisms. But on AmigaOS the nature of it would allow threading with direct on write access.

What I really mean is having a process that can run threads. Despite being in the same memory space, an AmigaOS process, which is a higher level than task, has it's own context. Despite using threading vernacular in the API, when a process creates a child to use as a thread, it's a totally different process. That's not a proper thread. A thread should belong to a parent process but an AmigaOS process has no concept of threading this way.

As an example, someone opens up a device, but wants to do the I/O work in another thread. They create a child process to do this. Why won't it work? Because the child process is a separate process with a separate context and own stack, memory and register setup. Had AmigaOS supported threads this would have worked, because the thread would have been running under the same process, but acting like a subroutine.

@Hypex

Just to nitpick, threads don't do copy on write. They operate in the same memory space as the process that created them. This is why you need mechanisms for explicit synchronization between them when they modify data. Forked processes do copy on write, when they modify any data they "inherited" from their parent memory space. It's implemented at a page access level since that can be implemented using MMU tricks for the most part and simply duplicating whole process space would be a monumental waste of resources.

Regardless of whether or not an exec Task or process has an explicit notion of a parent doesn't prevent it from being added in your own code. Years ago, I wrote my own multiplatform C++ framework that implemented something akin to the Java thread model based on Tasks. The exec Task user data pointer pointed into the Thread instance and it in turn managed parent child relationships. There was a root thread wrapper that represented the main entry point.

The AmigaOS 3.x implementation of the class is defined in a platform conditional include here https://github.com/0xABADCAFE/ancient-amigaos-exng2/blob/69c07dc1fcc215c81c2bf5ea7e25a83ca6a1d9e4/include/platforms/amigaos3_68k/systemlib/thread_native.hpp#L52

It worked very well on 3.x compiling with stormgcc and it's threadsafe.lib. Unfortunately it was impossible to get the OS4 implementation working because the compiler and toolchain were compiled in a single threaded way. Thus exceptions thrown in a child process ended up getting unwound in the parent, vice versa and it was a total disaster. One that I've never been able to rectify.

Last edited by Karlos on 03-May-2022 at 07:37 PM.

_________________
Doing stupid things for fun...

On a different note, as I'm sure everyone is tired of me harping on about 68K as the future...

Optimistic or pessimistic? I'll be honest. It depends. The Amiga was the creative computer for the creative mind. If the future is just the total stagnation/dead ending of NG due to an architectural ball and chain the key is at the bottom of the ocean for, a
or an anemic continuation of the platform solely for retrogaming purposes, yeah, that sucks in my opinion and gives me little reason to be hopeful. I can retrogame in UAE all day as it is. I don't need a honey I shrunk the Amiga toy for that. To have hope, I'd want to see people using it creatively. Making art. Making music. Making software. Creating stuff.

An Amiga without user creativity is like a broken pencil. Pointless.

_________________
Doing stupid things for fun...