ARM based Amiga?

[Tripitaka]

@Mrs Beanbag

No, there are things that are completely impossible to implement in Amiga in a clean way without breaking the Amiga compatibility, but would be totally possible if you would say "fcuk the old, only new stuff from here on" and start with a clean slate, and create a new API that will incorporate functions like these but without the ambition of running old Amiga apps (or system components for that matter).

Ahh... and that then begs the question:

Just what do we accept as "backwards compatibility"?
In the true sense of the term, you are of course correct. To some however (myself included) "invisible emulation" is a better way forward. This allows for a clean slate and my old games and apps.

[Mrs Beanbag]

No, there are things that are completely impossible to implement in Amiga in a clean way without breaking the Amiga compatibility

Such as...?

You could run UNIX on a classic Amiga so a multi-user environment isn't impossible.  It's not legacy hardware support that's holding us back here.  Obviously Microsoft and Apple have changed their APIs a lot over the years, Windows 3.1 software won't run on Windows 7 anymore, heck even XP software needs to run through an emulation layer.  If we need a new and radical AmigaOS then that's fine by me, but there's no reason it can't run old Amiga software through an emulation layer (although just running UAE on it is kind of cheating, it has to be integrated somehow so it doesn't "feel like" you're using an emulator).  This kind of emulation causes me no ideological problems.

If the hardware at least is backwards compatible then at worst we would be able to dual boot it.  As pointed out, an AGA chipset implementation would take up a small fraction of a modern chipset die, even if it included a 680x0 coprocessor.

[persia]

Have you actually tried to install Dos on an i7 or six core xeon?

I don't think it's impossible to have both backwards compatibility AND 21st century functionality.  Intel-based PCs manage it, after all.  You can still install Windows 3.1 on a modern PC, or even DOS.  You can still run software from 1981.  I'll repeat this: all modern PC graphics chips have legacy support for EGA graphics!  And let's face it, the continuous demand for backwards compatibility is the only real reason we're still using x86 chips at all, it's not because there are any inherent advantages to the instruction set.

[takemehomegrandma]

Such as...?

Such as true SMP and true MP, like I mentioned...

[tone007]

Have you actually tried to install Dos on an i7 or six core xeon?

I personally hadn't, but I had a few minutes...

i7:


Not going to try for audio drivers!

[Mrs Beanbag]

@persia: no but I've booted it from a USB stick on an Athlon II...

@takemehomegrandma: and what would be the barrier to that if we were running on a multi-core ARM processor?  I don't see why this should be the thing to break backwards-compatibility.  AmigaOS already multitasks, what difference does it make to an application exactly how it achieves it?

[commodorejohn]

Mrs Beanbag: It's not a hardware issue, it's a question of software architecture. The Amiga kernel was designed for hardware with no memory protection, so (in a less than perfectly forward-thinking move) they designed it to take advantage of the freedom that allows. Message-passing between processes, for instance, simply involves one process handing over a pointer to the message content to another and saying "have at it," and the other process freely accessing the first's memory. It's blazing fast, but it's also completely insecure.

Thus, there are actual software-architecture barriers to implementing something like memory protection - you'd have to figure out how to work around the existing API, and that's not necessarily easy. "Emulation layers" don't fix everything; it's entirely possible for one process to send another a whole handful of pointers and lots of wild, unprotected accessing of each others' data to take place without the OS knowing anything about it. (Granted, that's horrible coding practice, but when has that ever stopped anybody?) Thus, while I'm not convinced it's impossible, it most certainly wouldn't be easy.

[nicholas]

Such as...?

You could run UNIX on a classic Amiga so a multi-user environment isn't impossible.  It's not legacy hardware support that's holding us back here.  Obviously Microsoft and Apple have changed their APIs a lot over the years, Windows 3.1 software won't run on Windows 7 anymore, heck even XP software needs to run through an emulation layer.  If we need a new and radical AmigaOS then that's fine by me, but there's no reason it can't run old Amiga software through an emulation layer (although just running UAE on it is kind of cheating, it has to be integrated somehow so it doesn't "feel like" you're using an emulator).  This kind of emulation causes me no ideological problems.

If the hardware at least is backwards compatible then at worst we would be able to dual boot it.  As pointed out, an AGA chipset implementation would take up a small fraction of a modern chipset die, even if it included a 680x0 coprocessor.

Haiku with AROS hosted and 68k emulation would be my preferred system.

ZUNE port to Haiku bounty perhaps?

[nicholas]

Mrs Beanbag: It's not a hardware issue, it's a question of software architecture. The Amiga kernel was designed for hardware with no memory protection, so (in a less than perfectly forward-thinking move) they designed it to take advantage of the freedom that allows. Message-passing between processes, for instance, simply involves one process handing over a pointer to the message content to another and saying "have at it," and the other process freely accessing the first's memory. It's blazing fast, but it's also completely insecure.

Thus, there are actual software-architecture barriers to implementing something like memory protection - you'd have to figure out how to work around the existing API, and that's not necessarily easy. "Emulation layers" don't fix everything; it's entirely possible for one process to send another a whole handful of pointers and lots of wild, unprotected accessing of each others' data to take place without the OS knowing anything about it. (Granted, that's horrible coding practice, but when has that ever stopped anybody?) Thus, while I'm not convinced it's impossible, it most certainly wouldn't be easy.

Back in the day I had high hopes for the MorphOS Q/Box, but sadly it never came to fruition.

[nicholas]

I personally hadn't, but I had a few minutes...

i7:


Not going to try for audio drivers!

FreeDOS might have them.

[Thorham]

So then how the hell would it be an Amiga? :/

It wouldn't. Amigas ended with AGA, the other systems are alternative systems that users insist on holding back with AOS derivatives and dated software.

Amiga is a dead end in terms of hardware (and current old software), and users who want more need to realize that they need to break with Amigas and move on or they will be held back.

People who are truly interested in Amigas and still want more need to realize that it's only really going to happen with new software that's technically and functionally better than the, mostly old, programs that are available now. Chipset+680x0 can do much better, but only if high quality software is written from scratch for it (most of the time).

I think what he really meant was rather "backwards compatible"?

No, I mean that people who are not really interested in classic systems have to ditch Amigas and move on, after all, how good are things like AOS4 and MorphOs really? Or AROS? Why would you want to run AOS (or derivatives) on newer hardware anyway? So you can use old programs that are mostly superseded left and right? Not me, that's for sure (I can do that with my Amiga already and don't need a new machine for this).

Newer machines need to be their own platform instead of sticking to AOS (and Amiga itself needs to move away from this as well).

[SamuraiCrow]

Haiku is the best OS for the PC.  But there's no software for it.  I agree with Nicholas on the hosted AROS port to Haiku.  That would allow lots of AROS software to run on Haiku.

As for the 680x0, it's only as much of a dead end as the NatAmi requires it to be.  We'll be bringing up-to-date features to the N68050 but it will still be a softcore for quite a while.

@Thorham
While the 68020 is capable of better OS functionality, I doubt that it's going to help.  Most of the extended addressing modes are slow on the '020 and '030.  The '040 and '060 fixed that at the expense of some hardware compatibility particularly for floating point.

Also, memory protection would slow the real Classic Amigas to a crawl.  Believe me, it's not needed.  A better solution would be to build software on top of the PNaCl (Portable Native Client) sandbox that is going into the latest Chrome web browsers.  It will allow 97% runtime speed compared to standard native compiled code on x86 and still be secure.  All this without memory protection.

[Mrs Beanbag]

@commodorejohn:
I see your point.  There are two easy options that come to mind.

1. dual boot!  I think we're all agreed at least that compatibility with legacy hardware isn't the problem, but rather maintaining compatibility with legacy software in an up-to-date OS.

2. run all legacy (i.e. 68k) software in a shared virtual address space.  Legacy applications will still be able to interfere with each other, while new software can be completely secure.  Obviously this is something of a compromise, but how much does it really matter and how much effort is it worth?  In fact this should be possible even under Linux with a modified UAE, by which I mean, rewrite the functionality of AmigaOS in C and compile natively instead of emulating it, library calls could easily be trapped and Intuition windows and widgets mapped to X windows.  I might look into doing this in fact.

I don't see why SMP should be a problem.

@Thorham: all your objection seems to come from the way you define "Amiga".  You've defined it dead.  Other people define it differently.

I'd love to see a new Amiga personally because I think of Amiga not as a thing so much as an ideology.  Although it's also because I feel quite strongly that "the wrong side won", maybe it's just loyalty, but I can't accept that Amiga is dead and the prospect of a new market in ARM powered desktop/laptop/HTPCs makes me think that it has a real chance to get back on its feet again.  If you don't want a new Amiga then, well, just don't buy one, fair enough.

[commodorejohn]

2. run all legacy (i.e. 68k) software in a shared virtual address space.  Legacy applications will still be able to interfere with each other, while new software can be completely secure.

Not too bad of an idea, actually - that's how Classic on Mac OS X does it, it runs a whole OS9 instance as a user process. I don't know what kind of performance impact that might have on a 68k machine, though.

library calls could easily be trapped and Intuition windows and widgets mapped to X windows.  I might look into doing this in fact.

That'd be a lot of work - X is a completely different approach to windowing than Intuition, as I understand it. Still, it would be pretty cool.

I don't see why SMP should be a problem.

I don't know for certain, but I'd guess it has to do with the lack of memory protection, as you can't be sure that two simultaneous processes wouldn't be clashing over memory they're intending to take turns with.

[Tripitaka]

It wouldn't. Amigas ended with AGA, the other systems are alternative systems that users insist on holding back with AOS derivatives and dated software.

Amiga is a dead end in terms of hardware (and current old software), and users who want more need to realize that they need to break with Amigas and move on or they will be held back.

People who are truly interested in Amigas and still want more need to realize that it's only really going to happen with new software that's technically and functionally better than the, mostly old, programs that are available now. Chipset+680x0 can do much better, but only if high quality software is written from scratch for it (most of the time).
No, I mean that people who are not really interested in classic systems have to ditch Amigas and move on, after all, how good are things like AOS4 and MorphOs really? Or AROS? Why would you want to run AOS (or derivatives) on newer hardware anyway? So you can use old programs that are mostly superseded left and right? Not me, that's for sure (I can do that with my Amiga already and don't need a new machine for this).

Newer machines need to be their own platform instead of sticking to AOS (and Amiga itself needs to move away from this as well).

OK, I see what your saying and believe or not I think you make some valid points. If we take a cold hard look at what we have got we find the Amiga name itself is split (CUSA "Amiga", Ainc "Amiga", Hyperion "Amigaone"). The software is mostly old and out of date (not all of it of course, but mostly). The OS is fractured and the derivatives are just that, derivatives. The hardware is old but awesome or new(ish) but generic and not outstanding. So, yes, you have a point.

So what does that leave us with?

FPGA solutions for classic clones are fine as stand-ins for old hardware running classic software we may wish to run, so is UAE however.

Natami will be, to my mind at least, probably the greatest hobby machine ever but is at the end of the day just that, a hobby machine, Gunnar has made this quite blatantly obvious on many occasions. I must say I look forward to Natami, it looks like a great solution for classic Amiga and adds power to a classic piece of computer history that will, I don't doubt, create a vibrant hobby community enjoying new games and apps whilst learning about the real fun that computers should be.

If however your desire is for powerful and efficient modern hardware, designed from the ground up to be a joy to use for end users and programmers alike with strong graphics and multi-media capabilities as well as an OS that is logical in it's implementation in such a way that the end user actually knows what all those damn files are for (rather than the dll shotgun splatter that is a windows C: drive), then, my friend ...
....please let me know when you find it.

[Thorham]

@Thorham
While the 68020 is capable of better OS functionality, I doubt that it's going to help.  Most of the extended addressing modes are slow on the '020 and '030.  The '040 and '060 fixed that at the expense of some hardware compatibility particularly for floating point.

If Amigas are going to go anywhere, then a new OS+software is needed.

Also, memory protection would slow the real Classic Amigas to a crawl.

Depends on how it's implemented ('030 MMUs can do some neat things), and whether or not it's used for everything (it would be great to turn it on just for programming).

A better solution would be to build software on top of the PNaCl (Portable Native Client) sandbox that is going into the latest Chrome web browsers.

Two problems: 1) This doesn't run on Amigas, 2) What fun is it to write for a browser?

@Thorham: all your objection seems to come from the way you define "Amiga".  You've defined it dead. Other people define it differently.

My definition of Amiga is hardware, namely chipset+680x0, and sadly this is dead until someone buys or licenses Amiga hardware and starts building the things, and that's most likely not going to happen

I'd love to see a new Amiga personally because I think of Amiga not as a thing so much as an ideology.  Although it's also because I feel quite strongly that "the wrong side won", maybe it's just loyalty, but I can't accept that Amiga is dead and the prospect of a new market in ARM powered desktop/laptop/HTPCs makes me think that it has a real chance to get back on its feet again.  If you don't want a new Amiga then, well, just don't buy one, fair enough.

My problem is this: Take the C64 and it's OS, now the Amiga comes along, and they put a C64 like OS on it. This is what's happening with the new alternative machines that superseded Amiga hardware. Those alternative machines are stuck in the past if they stick to the Amiga legacy, instead of just running an emulator.

Want an Amiga? Use/get one. Want newer hardware? Then move on, and do better (because there's room for a lot of improvement that won't happen by sticking to AOS).

And yes, I want a new Amiga, but no one will make one I'll probably get an alternative machine after Amigas die physically (hopefully that won't be for another 10+ years...).