Amiga OS 4.x, is the dream over?

[amigang]

Here my two cent, its a real shame, there is something about using an Amiga that make me keep coming back and wishing i had a modern version.

Yes the classic always fun, and will always have many memories of the old gaming days and creativity I had on it. I respect everyone view that feels it should be just that and thats it, but I really would love to dump windows 10 and have my Amiga.

OS4 is great and it such a shame in way that its feel 80% of the way I want a NG Amiga to feel. its that last 20%, where I do feel I still hit a few too many grime reaper/lockups and just too many task I now expect to do on a modern platform, my poor X1000 just cant do or it not fun to do (it pains me to say that) but due to the lack of support/development, and it no ones fault really, PPC seemed a wise choice back in 1999, but so many factors and things happened that missed up Amiga next step.

I do love OS4/x1000 it my main go to Amiga now. I dont really use my classic , just emulate the odd game on my PC, started playing around with Amikit, which is pretty impressive and shows how far the classic AmigaOS can be pushed.

but basically my point is, OS4 and the whole current AmigaNG market is not great but it what we got and for the size of the community and money that has been involved in trying to push Amiga forward its seriously impressive what we have achieved, would I like it to be better, YES, do I think the future bright for it....not sure..but I can still have fun on it and at the end of the day that all that matters.

[manossg]

Originally Posted by Thomas Richter

It is a bit contradiction in terms to modernize a retro machine.

Why? Is Spectrum Next a contradiction?

Originally Posted by Thomas Richter

Second, I am not a particular fan of ignoring good development practice and ignore guidelines for Amiga hardware development.

Care to elaborate on your views on good development practice and your own guidelines for Amiga hardware development? I am honestly curious.

[Retro1234]

I think that hyperion themselves couldn't create a decent Web-browser was the final straw.

[Samurai_Crow]

Originally Posted by manossg

Care to elaborate on your views on good development practice and your own guidelines for Amiga hardware development? I am honestly curious.

In order to free up space in the 68k instruction set for AMMX, Gunnar had to eliminate some seldom-used features such as the ability to map subroutines into the instruction set. The performance of these subroutines were slower than just using an absolute jump to them IIRC.

[Thomas Richter]

Originally Posted by manossg

Why? Is Spectrum Next a contradiction?

I do not know, so I cannot really judge. Sorry.

Originally Posted by manossg

Care to elaborate on your views on good development practice and your own guidelines for Amiga hardware development? I am honestly curious.

Certainly, but beware that this is from what I know, and I may not be up to date.



Expansions enter the Amiga world through AutoConf, a pre-cursor of the PCI-self-announcement protocol. This allows to adjust software, drivers... to a particular board, and there is a whole protocol in the Os that allows loading and booting from such devices, or support software, or adding RAM to the board.



Vampire does not use autoconf.


Address space: The address space of the Amiga is documented, with the F0-space reserved for debug-ROMs, though hijacked at its time by P5 for their 68060 boards. Bad enough. The E8-space is reserved for Kickstart-expansions, allocated by the CD32.


AFAIK, the vampire does not use the F0-space, but the E8 space for its expansions, making 1MB roms impossible to realize. The reason for this is that the Atari ST has its IDE interface sitting in the F0 area, conflicting with the Amiga "debug ROM space". An alternative would be bank-switching for ST emulation, a truly old technology.



To my very knowledge, vampire requires a patched kickstart to scan its own private modules in the E8 space - instead of the F0 space. Requiring to patch up kickstarts is certainly bad development practise.


To the latest of my knowledge, the vampire makes use of some already-defined opcodes of the mot-family to place its own extensions there. Certainly less-used opcodes, and even ones that make no sense on the Amiga, though this renders a CPU-independent disassembler (for example) impossible. There would be 6 co-processor opcode slots, each of which with many many opcode slots available for such extensions. Not following the design principles of Motorola is something I consider bad practise.


To my latest knowledge, the vampire uses a different supervisor model, in particular as far as memory attributes and memory protection goes. That is the whole world of the MMU, a very useful tool for debugging and professional development. This requires a vendor of system tools to re-develop its tools for this particular hardware, instead of just requiring a small system library driver which would make all these tools available instantaneously. I consider deriving from a well-working, well-tested system model a bad development practise.


To my knowledge, vampire provided a driver for their chunky graphics modes in a very questionable way, probably debatable, through first checking with their authors, and then bypassing them instead of arranging for a compensation for their work. I consider this bad teamplay.


But then again, these stories are old. I do not critize the technical work of the folks, it is amazing, but there is too much foul play in the game I do not feel comfortable with.

[manossg]

Originally Posted by Samurai_Crow

In order to free up space in the 68k instruction set for AMMX, Gunnar had to eliminate some seldom-used features such as the ability to map subroutines into the instruction set. The performance of these subroutines were slower than just using an absolute jump to them IIRC.

Not sure I understand how these affect day-to-day performance, but thanks, it is good to know.

[EDIT]
@Thomas Richter

Thanks, man, a good read.

[meynaf]

Originally Posted by Samurai_Crow

In order to free up space in the 68k instruction set for AMMX, Gunnar had to eliminate some seldom-used features such as the ability to map subroutines into the instruction set. The performance of these subroutines were slower than just using an absolute jump to them IIRC.

AMMX uses coprocessor space, so it looks safe to me. The placement of the other extensions, however, is more questionable.

Originally Posted by Thomas Richter

To the latest of my knowledge, the vampire makes use of some already-defined opcodes of the mot-family to place its own extensions there. Certainly less-used opcodes, and even ones that make no sense on the Amiga, though this renders a CPU-independent disassembler (for example) impossible. There would be 6 co-processor opcode slots, each of which with many many opcode slots available for such extensions. Not following the design principles of Motorola is something I consider bad practise.

If things haven't changed from what i know, it is CALLM space that was reused. Nevertheless a bad practice.

[gulliver]

Originally Posted by Retro1234

I think that hyperion themselves couldn't create a decent Web-browser was the final straw.

I believe you hit the nail on this one:

To have any platform become your daily driver, while you can have several lacking areas, what you must have fully working with all bells and whistles is a web browser. You cannot workaround that today.

AmigaOS 3.2 follows another much different path and goals than OS4. And who can blame them for getting impatient due to the current hardware and software scenario they have been facing? Some OS4 users and devs will probably migrate to 68k, after all, it was their first love.

But also many of them will stay chasing their dream of a Next Generation machine to compete with modern PCs no matter what.

I do not share their dream, nor their vision. But I can certainly respect their will and effort to pursue their goals.

Can you judge them for following their dreams?

[utri007]

Amiga 68k development primary goal should be maintain compatibility. Problem is both Vampire and OS3.1.4/3.2.

If my old boot floppies doesn't work they are useless to me. I have bought OS3.1.4 because of SCSI.device, cdfilesystem and some other components, price is not bad after all.

Both all equally dangerous, both will fragment scene. Make our ecosystem smaller, like MorphOS and Aros did for Amiga OS4.

But there is no comeback for Amiga with any form, these are hobbyists machines and OS, so they will never die.

[Samurai_Crow]

@utri007

MorphOS and AROS both came out BEFORE OS4. Which one did fragmentation to WHOM?

[daxb]

You just need to adapt your old boot floppies if you want to run them with new OS. Should be an easy task. If you don't like to, stick to the old. With a floppy only system you are limited in some ways. I have ~30 years old boot floppies. They become already obsolete when I switched to A1200.

[utri007]

Originally Posted by daxb

You just need to adapt your old boot floppies if you want to run them with new OS. Should be an easy task. If you don't like to, stick to the old. With a floppy only system you are limited in some ways. I have ~30 years old boot floppies. They become already obsolete when I switched to A1200.

I don't have a single floppy drive. You missed my point. I give a hint, what else doesn't work than old boot floppies?

[Thomas Richter]

Originally Posted by utri007

Amiga 68k development primary goal should be maintain compatibility. Problem is both Vampire and OS3.1.4/3.2.

If my old boot floppies doesn't work they are useless to me. I have bought OS3.1.4 because of SCSI.device, cdfilesystem and some other components, price is not bad after all.

Well, as said in other places: Updating such an old operating system requires to make compromizes as assumptions made back then are no longer valid today.

For example, you do not get floppies these days anymore. It was actually a big problem finding some to ship Os updates. Thus, one decision for 3.1.4 was to require users to have a hard disk, and keep some system components there, and to prepare the Os to find system components there instead.

Booting 3.1.4 from floppy is possible, but not much fun, and not the envisioned modus operandi.

[utri007]

Originally Posted by Thomas Richter

Well, as said in other places: Updating such an old operating system requires to make compromizes as assumptions made back then are no longer valid today.

For example, you do not get floppies these days anymore. It was actually a big problem finding some to ship Os updates. Thus, one decision for 3.1.4 was to require users to have a hard disk, and keep some system components there, and to prepare the Os to find system components there instead.

Booting 3.1.4 from floppy is possible, but not much fun, and not the envisioned modus operandi.

Well I respect you decision I woun't start "kolling" about this. Just hope that you could keep some rommable workbench and icon libraries on mind. Maybe add support 1mb rom for A1200, because it is after all supported by hardware.

[Fastdruid]

Originally Posted by coldacid

Given that you can't get an X5000 for less than three grand these days (after the conversion from GBP) it shouldn't be a blip on anyone's radar.

Honestly I still feel like a big part of the reason that OS4 has been a flop is because it's far more limited in where it can run. Unlike m68k hardware which is practically ubiquitous, the only options for OS4 are emulation, classics with a CyberstormPPC, or a small set of boards blessed by the touch of Hyperion. There's plenty of PPC hardware out there it could have run on, but it would have meant less control (and/or money) for Ben Hermans.

Y'know what, I'm just going to blame him for it.

Personally I feel it's a chicken and egg scenario.

I *LIKE* OS4 but without any modern software (including a web browser) it's little more than a hobby/plaything. Which would be fine if the hardware was peanuts but it's top of the range modern hardware prices for what is 10-20 year old design hardware.

I can justify the £3k or so for a PC that I'm going to *use* (the current one I'm on now I've probably spent that on) or even a few 100 for a second hand Mac and run MorphOS on it but I can't justify that as a plaything with little more than just the OS on it just because it's "nice" and the market for people who *are* prepared to drop £3k on something like that is tiny.

So with it being far too expensive and not running the games (which can be done far cheaper anyway) it needs modern software to stop it being _purely_ a hobby plaything...but no one is going to develop it if there is no user base for it and without the software there is no market for it and so no user base etc etc etc.

Equally no user base, no cheap hardware, no cheap hardware no user base. Round and round we go.

If OTOH for example it could run everything 68k (in emulation) as well as doing say something akin to WINE and run Windows or Linux software then even at £3k it would be (semi) justifiable. Equally if you could pick one up for lets say £3-£400 (as a system) with 2-5 year old spec hardware rather than 10-20 year old then it's cheap and powerful enough and the lack of software doesn't matter.

PPC was in hindsight a mistake. It appeared at the time to be a good way to move forward from 68k but progress was too slow onto it and by wanting to be "NG" it had to keep up with the Joneses for processors and that was only going to happen while PowerPC itself was widely used by a significant market. When Apple decided to go x86 in 2005 that was it. That means in short that before AmigaOS 4 was even released, it was already doomed.

Anyway you slice it, AmigaOS 4 is little more than a tiny niche product and without AmigaOS doing an Apple and going x86 or ARM (and we know why because of the whole legal issues it can't) there it will stay.

Which is a crying shame because I *like* it, I really do. I just can't see myself ever having any hardware to run it.

[jPV]

Originally Posted by utri007

Both all equally dangerous, both will fragment scene. Make our ecosystem smaller, like MorphOS and Aros did for Amiga OS4.

Originally Posted by Samurai_Crow

@utri007
MorphOS and AROS both came out BEFORE OS4. Which one did fragmentation to WHOM?

Yeah, and it was a real option that MorphOS would have become the official OS4 and we would have been more united, but as usual some egos clashed and they decided to fragment the market by creating the current OS4 from scratch.

[Thomas Richter]

Originally Posted by Fastdruid

Personally I feel it's a chicken and egg scenario.

No, it is a "missed market" understanding. The way how Os 4 is run is IMHO under the assumption of a home-computer market we had in the 1980: Provide the hardware, an Os, some documentation and users care about the rest.


This model does not work anymore. It died around 1990. Any attempt to base a business on this model is doomed to fail, and PPC and AmigaNG was doomed to fail from the very beginning. Not that I did not say that in 1990...

I really do not see a niche for Os 4 at all - not back then, not today. It simply does not address any particular need a customer may have. If I need the software and the performance, then a PC will do, including emulation. If I need retro-computing, the 68K model will do. If I need "homebrew computing", the raspi will do more than fine (at a fraction of the price).



The problem Os 4 is attempting to solve is just unclear to me.

[Tigerskunk]

Originally Posted by Thomas Richter

If I need the software and the performance, then a PC will do, including emulation. If I need retro-computing, the 68K model will do. If I need "homebrew computing", the raspi will do more than fine (at a fraction of the price).

This.
So much..

Either something is nostalgic, or it is practical.

OS4 is neither.

[jPV]

Originally Posted by Thomas Richter

I really do not see a niche for Os 4 at all - not back then, not today. It simply does not address any particular need a customer may have. If I need the software and the performance, then a PC will do, including emulation. If I need retro-computing, the 68K model will do. If I need "homebrew computing", the raspi will do more than fine (at a fraction of the price).

The problem Os 4 is attempting to solve is just unclear to me.

It's for us who did expand our classic Amigas to the top around the millennium and tried to use them as our main operating systems as long as possible, simply because we love how the OS and certain software works on it and how we can do our workflow on it.

NG did step in when we had reached the limits on both hardware and software on our classic setups. It allowed us to use much more modern and powerful hardware with the way we like to use computers. Also software and OS developed much further from OS3 but keeping the underlying philosophy and even compatibility. On OS3 side people tried to patch and hack new features to the system, but it's always been unreliable and on NG systems you get them implemented in clean way to the OS itself and they became much more stable to use.

PC just didn't, and still don't, do it for me, because I love to use things in Amiga way and I don't want to settle down to the "wil do" situation, I want the same enjoyment I got with classics. Classic Amigas are still falling apart because of their age and they still can't be expanded to the NG level of performance even with Vampires etc. And even if there would be some powerful enough standalone Vampire like "classic" machine, it would still lack on OS side compared to NG operating systems. It would just be too big step backwards when, for example, MorphOS has been developing forward for over 20 years now. This same reason is why I don't see emulating OS3 as a viable option for me either.

I chose MorphOS as my continuation for the classic Amiga, because I think it's the best for my needs, but if it wouldn't be available, I would use OS4.

I don't know why it is so hard to understand and respect this viewpoint for many of you, who did see the Amiga world desperatively expanding their systems based on stone age motherboards, and how it seems to be so unclear that it should be asked or even evangelized against? Can't we just respect and appreciate that we have options even if you don't happen to need it yourself? Why are we shooting each other down in our own hobby? Someone else might think different, but doesn't even try the other options because of this kind of talk... and getting interest towards any of the incarnations would finally help all the others too. After all, the situation is what it is, and we can't change the history and make anything vanish by talks.

[Thomas Richter]

Originally Posted by jPV

NG did step in when we had reached the limits on both hardware and software on our classic setups.

Software? Which software? Os 4 does not have a large software basis.

Originally Posted by jPV

It allowed us to use much more modern and powerful hardware with the way we like to use computers.

I do not understand. How do you like to use your computer? Look at the workbench twice a day? I am sorry, but without a large software basis, I do not quite see the point in using a computer.

Originally Posted by jPV

Also software and OS developed much further from OS3 but keeping the underlying philosophy and even compatibility.

Which compatibility? Can I run 68K software? Oh, in emulation, sure, but the PC can do that either, just faster, and a lot cheaper.

Originally Posted by jPV

On OS3 side people tried to patch and hack new features to the system, but it's always been unreliable and on NG systems you get them implemented in clean way to the OS itself and they became much more stable to use.

I wouldn't know that 3.1.4 was an unreliable patch job.

Originally Posted by jPV

PC just didn't, and still don't, do it for me, because I love to use things in Amiga way and I don't want to settle down to the "wil do" situation, I want the same enjoyment I got with classics.

Do "which things"? The things I do can be done on the PC, faster, better, cheaper. I mean productive work. Not retro-computing. That is what the 68K line is about.

Originally Posted by jPV

Classic Amigas are still falling apart because of their age and they still can't be expanded to the NG level of performance even with Vampires etc.

Though that is not even the point. This is a museum piece, a historic artefact.

Originally Posted by jPV

And even if there would be some powerful enough standalone Vampire like "classic" machine, it would still lack on OS side compared to NG operating systems.

Why would it need to? If I would need a modern Os, it would surely not be anything Amiga-like. Not even remotely. There are too many constructional errors in it to allow future expansion. Single-processing, no memory protection, no resource management, nothing that makes this thing even remotely interesting from an Os point of view. Neither NG nor 68K, but 68K is at least an interersting historic artefact with a huge software basis. NG is neither modern, nor useful, nor historic. So what is it good for? I do simply not understand the point.


Yes, appreciated that lots of work hours went into this, of course. But what for?

Originally Posted by jPV

It would just be too big step backwards when, for example, MorphOS has been developing forward for over 20 years now. This same reason is why I don't see emulating OS3 as a viable option for me either.

I couldn't care less about Morphos either. Same problem. It has no historical value, and neither a practical value.

Originally Posted by jPV

I don't know why it is so hard to understand and respect this viewpoint for many of you, who did see the Amiga world desperatively expanding their systems based on stone age motherboards, and how it seems to be so unclear that it should be asked or even evangelized against?

It is so hard because I see no point in a hardware basis with no software basis, nor particuarly interesting as historical value, nor much of a practical value.


Sure, it's a free world, do what you like to do. I just do not see a way how to share this enthusiasm about... about what exactly?