@Karlos

Sure running AOS hosted and then using apps on the host side is an option.
If you have come to the conclusion that AOS is 100% obsolete.

Why you would need an "uber" fast 68k in that scenario is beyond me.

AxRT is like so many things else a nice experiment that may or may not yield bigger results in the future but in it's current state is simple irrelevant for anything Amiga (or not Amiga).

Is my G5/MorphOS combo 10% obsolete? 50%? Or maybe 99%? For sure not 100% as I can use it for everyday stuff if I really want.

Of course it will get closer to that 100% with every year it doesn't get neither a structural update or better HW. And even if it does get both it might still do.
It just ain't there yet.

_________________
- We don't need good ideas, we haven't run out on bad ones yet
- blame Canada

@Kronos

I don't know the answer to that, I'm not the one making a case for degrees of obsolescence based on speed. It would be highly subjective.

My point is, and I hope this has been obvious from the start, that *architecturally* all AmigaOS / compatibles based on 32-bit single unprotected address space and single core processing models are equally obsolete because they are each fundamentally the same. It's true that the NG derivatives have improved APIs and peripheral support, but those are incremental improvements that could be implemented across the board, even if nobody has. The same applies to the software they *can* run, even if it's completely insanely impractical, someone *could* build a more modern web browser that runs on an 020+ with enough memory and it might even be usable on simple sites on actual silicon (68060/RTG probably). It might even be useful under emulation.

However, that browser won't have many important features you take for granted elsewhere - process isolation with separate memory maps, JIT execution for JS, wasm support, etc. You won't get gigacage style mitigation against JS out of bounds / overflows and other mitigations. These all require 64-bit addressing to be feasible. A browser that runs under AmigaOS should be trivially easy to exploit by someone that knows about it - they may even be able to scan the entire memory by exploiting out of bounds accessed. I don't know what sort of mitigation MorphOS had against this but I don't imagine it's much better for anything inside the A box.

As for why I want to run it at all, especially a faster iteration, well that's simple. It's fun and I enjoy it.

Isn't that why we all do?

Last edited by Karlos on 14-Mar-2025 at 07:38 PM.

_________________
Doing stupid things for fun...

@Karlos

I would say it like this

in a way currently everything amiga related is retro

but only 68k users admit it

@Karlos

AROS already runs on X86, there is even a 64bit version supporting SMP. Problems are drivers and software. And there all ends. You cannot overcome that without big investments.

@Kronos

When I made a benchmark at least winuae on a halfway fast PC beats most of the PPC options running native apps. So PPC is retro too, just a little less than 68k

@Karlos

Quote:

Karlos wrote: @Kronos I don't know the answer to that, I'm not the one making a case for degrees of obsolescence based on speed. It would be highly subjective.

Sure it's objective, some people may say that any HW that can't run the latest AAA game at 60FPS+ is obsolete.
Others will set the point at the moment when they stop receiving automatic updates for the OS and you'll even find some who consider anything 2 generations/years old obsolete.

And then you have those you will find ways to put the latest OS on HW 10 years out of support or install Linux on stuff that is even older.

Obsolete is something that can't be used anymore for it's intended purpose but what you define as it's purpose and what level of compromise you might consider useable is highly subjective.

What isn't subjective is that a system with better SW and/or higher speeds will push that line further out, no matter where you draw it.

_________________
- We don't need good ideas, we haven't run out on bad ones yet
- blame Canada

@Kronos

less obsolet as Morphos I would say

@Kronos

On AROS 32bit (X86) there is a updated browser

And you obviously do not know or understand what AxRT is

@OlafS25

The AROS Odyssey is behind/older Wayfarer but runs on faster HW. So I'd guess it will show fewer sites faster a compromise I don't want or need to make at this time.

I do know what AxRT is but I don't see the point in using the oddball AROS app, somewhat integrated but still kinda wonky when using Linux which has all those apps in more current versions (or better alternative in the case of FinalWriter).

In that way everything Amiga is indeed obsolete as everything you can do with can be done easier&better with Win/OSX/Linux.

As I said everybody has to draw the line somewhere, and for me it has to feel native (read Amithlon or Pimiga would qualify) and I have to be able to do basic stuff like checking email, or sending of a Google search without turning to another system (wether it's the host or another computer).

If AxRT or a potential MorphOS4.0/AOS5 ever gets to the point where it all feels "Amiga" consistently and has all bells and whistles that are to be expected in this day and age, then I will use that.
But right now we only get various levels of halfbacked and I see no point switching from one to another.

_________________
- We don't need good ideas, we haven't run out on bad ones yet
- blame Canada

@Kronos

all platforms are halfbaked and lack software so indeed changing is not needed

but finally all are retro in the sense of not on the same level of the big platforms

in my view AxRT is the only realistic plan to get something useable

part of this is a amiga desktop as real linux desktop (here scalos) and a window system making linux apps look and feel like amiga apps

https://www.youtube.com/watch?v=BFKy-8PfWmk

we will see what future brings

Last edited by OlafS25 on 14-Mar-2025 at 08:52 PM.
Last edited by OlafS25 on 14-Mar-2025 at 08:51 PM.

@Kronos

Quote:

Kronos wrote: @Karlos Quote: Karlos wrote: Hang on there squire. Name one single architectural leap that MorphOS (or OS4) has made over 68K that was *only possible* because they run on PPC.... And here I was thinking you were a native in English with a basic grip on logic.... But to make it simple: I did state that PPC/ARM/x86 can't revive the Amiga. But it should also be clear that the same reasons apply for anything 68k. Real, fake or emulated.

At least 68k can run all Amiga software, so it's in a much better position compared to all post-Amiga OSes (AROS, MorphOS, AmigaOS).
Quote:

@matthey Real 68k (100MHz 060) are at levels that would have been o.k. in 1997. Fake 68k (anything FPGA) are somewhere around 1999. Emulated 68k is in 2001 on ARM and maybe 2005 on x86. (MorphOS/OS4 are on 2000-2005 levels depending on the HW used) Anything less ancient would require 64bit and proper SMP both would require throwing out the whole existing stack of SW (or at least hiding it in a sandbox). At which point going for pie-in-the-sky made up super 68k wouldn't make any sense. Or in short we are doomed anyways you may just as well enjoy the ride.

Then you can embrace AROS. As Olaf already reported, it supports both 64-bit and SMP: something which neither AmigaOS nor MorphOS can do (and the original Amiga o.s. as well).


@agami

Quote:

agami wrote: @Kronos Quote: Kronos wrote: No I don't think anything PPC/ARM/x86 can MakeAmigaGreatAgain I agree, the only thing that can make a new Amiga thing greater than or as great as the original Amiga thing (however which way you measure Amiga greatness), is with RISC V. Here are the steps to new Amiga greatness: 1. Get all previous 68k software running on globally affordable hardware. Without new 68k CPUs fitting this requirement, we resort to 68k JIT on ARM32. 2. Gradually rewrite as much of the OS (AROS) to be ARM native. Ship of Theseus. 3. Re-architect AROS to be a multi-core 64-bit ARM64 little-endian first class citizen. Sandbox the legacy. 4. Port ARM native AROS running ARM apps and sandboxed legacy 68k to RISC V. V. Design our own RISC V cores to include a high performance 68040+ core.

That's a complete non-sense, as kolla already reported:
- ARM32 is a non-sense (even ARM completely stopped developing it);
- rewriting AROS as well (a port is sufficient, since almost all code can be retained: it's how AROS is. For whom knows it);
- AROS is already 64-bit and SMP-capable, and it "only" needs an ARM64 port. And it also has a sandbox for the legacy stuff (Janos);
- it's a total non-sense porting AROS from ARM to RISC-V: you "just" (!) need to port AROS to RISC-V;
- the last I don't even comment..

Many ideas and A LOT of confusion...

BTW, RISC-V is a very weak architecture. If I had to chose a LD/ST (since there are no "RISC" architectures), then AArch64/ARM64 is the best one.
I assume that you decided for RISC-V only because it's the buzz word of the time, without a prior technical analysis: a "good" way for selecting an ISA...


@Karlos

Quote:

Karlos wrote: @Kronos Has it crossed your mind that Uber 68K users may not be as bothered to have a modern browser - particularly if they are running on an emulation that already has one on the host. You know, a not obsolete one, one that supports modern standards and technology and is kept up to date by a large development team. Don't get me wrong, I do think a browser more akin to Wayfarer would be a value add, but it's not going to up to the same standard as any major browser that runs on an up to date 64-bit SMP OS. But this is where AxRT comes in... The price you pay for having all the modernity you want is that you can't directly run old binaries though I suspect at some point more deeply integrated emulation will make that problem redundant, even if it's just UAE in the background running a 3.x image.

AxRT is something different: it's Linux which is running, and not an Amiga / post-Amiga OS.

It's nice solution, but it doesn't match with the current discussion.

@cdimauro

Regarding AxRT, I think you missed the point I was making. AxRT delivers the promise that NG originally made with 64-bit, SMP and support for modern software, a modern OS. The price paid for this is that it's an entirely different OS. Linux in this case, but the point is it could be any other 64-bit SMP capable OS, the end result is the same: you have no binary compatibility outside of some form of emulation/sandbox.

AmigaOS 3, 4, MorphOS as they exist today are completely stuck in the past. Personally, I don't have an issue with this (I love 68K and PPC is fine) *except* when people make retarded arguments that PPC Amiga/like are somehow less obsolete than 68K when they are no closer in any meaningful way to modernity. They are merely faster at being obsolete 32-bit monoprocessor systems. The speed advantage they have doesn't qualify since you can strain that same advantage over 68K silicon with UAE.

People who make these arguments are as stuck in the past as the architecture itself IMNSHO.

_________________
Doing stupid things for fun...

@Karlos

Quote:

Karlos wrote: @cdimauro Regarding AxRT, I think you missed the point I was making. AxRT delivers the promise that NG originally made with 64-bit, SMP and support for modern software, a modern OS. The price paid for this is that it's an entirely different OS. Linux in this case, but the point is it could be any other 64-bit SMP capable OS, the end result is the same: you have no binary compatibility outside of some form of emulation/sandbox.

That's granted, but it's an hosted solution, which I personally don't consider when comparing OSes like the discussed ones.

They are very nice and comfortable, and that's the reason why I use them. But not native.
Quote:

AmigaOS 3, 4, MorphOS as they exist today are completely stuck in the past. Personally, I don't have an issue with this (I love 68K and PPC is fine) *except* when people make retarded arguments that PPC Amiga/like are somehow less obsolete than 68K when they are no closer in any meaningful way to modernity. They are merely faster at being obsolete 32-bit monoprocessor systems. The speed advantage they have doesn't qualify since you can strain that same advantage over 68K silicon with UAE. People who make these arguments are as stuck in the past as the architecture itself IMNSHO.

I think that either they aren't technical people (so, they don't understand the technical aspects), or are simply (!) blind fanaticals.

I've already written an article about the topic, and shared here a well: there's really nothing of all post-Amiga OSes which can be classified as "NG", with the exception of AROS which offers 64-bits (but AFAIK it's still using 4kB pages, which is useless without virtual memory and/or memory protection. Better to use 1GB pages, because it basically zeros the overhead of page walking, improving the performance) and SMP (but there are still problems).

Non technical people don't know and probably don't get the difference between a PORT/REWRITING and a brand new thing (which removes some old limits and/or introduces new features which were NOT possible with the old OS).

Blind fanaticals... I don't need to comment.

Retarded... well, there are many of them and unfortunately we can do nothing here: Nature was a very bad step-mother...

@cdimauro

Quote:

cdimauro wrote: nothing of all post-Amiga OSes which can be classified as "NG",

Guess what "Next Generation" means just that and nothing more.

The C64 was that VIC20-NG even if it just did more of the same thing.

A500/2000 were A1000-NG simply because of the fact that they followed it.

Whether you call A600/CDTV/A3000 the NG for those is up to you.

A1200/4000/CD32 were again "NG" when they came out.

So yes, NG Amigas are NG, simply because they came later and different is some ways.

Nothing technical about, and surely nothing that could be pin pointed to specific features. Them failing on delivering what they promised or failing altogether? Irrelevant.

_________________
- We don't need good ideas, we haven't run out on bad ones yet
- blame Canada

@Kronos

Quote:

Kronos wrote: @cdimauro Quote: cdimauro wrote: nothing of all post-Amiga OSes which can be classified as "NG", Guess what "Next Generation" means just that and nothing more. The C64 was that VIC20-NG even if it just did more of the same thing. A500/2000 were A1000-NG simply because of the fact that they followed it. Whether you call A600/CDTV/A3000 the NG for those is up to you. A1200/4000/CD32 were again "NG" when they came out. So yes, NG Amigas are NG, simply because they came later and different is some ways. Nothing technical about, and surely nothing that could be pin pointed to specific features. Them failing on delivering what they promised or failing altogether? Irrelevant.

Nobody called a C64 as VIC20-NG. I haven't called my C128 as NG after my Plus/4. And I haven't called my 1200 "NG" after that I've exchange my 2000 for it.

Those are simply "successors" of the previous model, which not necessarely have to share something or be a continuation of.

NG belongs only to Star Trek.

On the Post-Amiga land it's only a label under which there is still the same obsolete technology with all the same limits of the original one (with the exception of AROS, that at least brought something new AND desirable).
As I've written, those are PORTS/REWRITINGS. And that's how they should be called, instead of inventing things only to stimulate the fantasies of the customers to let them think that it's "the next big thing".