@pavlor: I also talked about NT in my posts about the argument, and I've showed why Windows was an o.s. with a good design, which allowed to have many drastic changes in the implementation, but still keeping a robust API compatibility.

And that's not something which is restricted to Windows: many o.ses had a good design.

The Amiga o.s. was an "hack & dirty" o.s. which lacks the necessary abstraction that would have allowed it to be quickly ported to any other platform and introducing more robustness too, without giving up its design (and APIs).

But, again, I repeat: I've talked a lot about it in the past.

@cdimauro

I took nearly 7 years for NT to become useable for general public (Windows 2000). That was with unlimited resources of Microsoft, gargantuan 3rd party market and broad base of compatible applications. With such power, even I would be able to port AmigaOS to x64 in two years time.

Quote:

cdimauro wrote: The Amiga o.s. was an "hack & dirty" o.s. which lacks the necessary abstraction that would have allowed it to be quickly ported to any other platform and introducing more robustness too, without giving up its design (and APIs).

I think "hack and dirty" is unfair to the AmigaOS. You are overlooking the good and exaggerating the bad.

+ efficient preemptive multitasking
+ practically 32 bit clean addressing and extensive dynamic addressing
+ use of C provided performance and portability
+ modularity and code re-use rarely achieved today
+ high and low level APIs

+/- performance and responsive due to user level libraries, pointer based message passing, efficient interrupt handling and DMA, etc.

- not enough abstraction and separation of functionality in some cases
- not enough forethought into later/optional memory protection, resource tracking and security
- AmigaDOS was inconsistent with the rest of the AmigaOS
- there should have been C sources for all assembler optimized code

The AmigaOS has aged better and with fewer changes than DOS with Windows, MacOS, Unix, RISC OS, Atari TOS with GEM, etc. Most of the operating systems of similar age were either replaced outright and/or mostly rewritten over time to become what they are today. The AmigaOS has not evolved as much because of an extreme focus on compatibility and lack of investment. Many of the negatives above have been partially dealt with but some changes are a tradeoff which will reduce performance and compatibility with minimal benefits.

@pavlor

Quote:

pavlor wrote: @cdimauro I took nearly 7 years for NT to become useable for general public (Windows 2000). That was with unlimited resources of Microsoft, gargantuan 3rd party market and broad base of compatible applications. With such power, even I would be able to port AmigaOS to x64 in two years time.

This is a completely different topic, pavlor.

Anyway, look at AROS: how many resources had the developers to build it from scratch and port to x86, x64, PowerPC, 68K, and ARM (works only from host now, but at least is running)?

@matthey

Quote:

matthey wrote: Quote: cdimauro wrote: The Amiga o.s. was an "hack & dirty" o.s. which lacks the necessary abstraction that would have allowed it to be quickly ported to any other platform and introducing more robustness too, without giving up its design (and APIs). I think "hack and dirty" is unfair to the AmigaOS.

Well, they needed to have something working ASAP, so they completely ignored the necessary abstraction. That's why "hack and dirty".
Quote:

You are overlooking the good and exaggerating the bad.

I was only talking about the bad, because that's the reason why the Amiga o.s. is so fragile and prevented to introduce modern features while keeping a good compatibility with APIs.
Quote:

+ practically 32 bit clean addressing

The address space is 31 bit effective, because of absolutely absurd decisions with some API.

And there should be some API (which I don't remember now) which used 24 bits.
Quote:

and extensive dynamic addressing

What do you mean with this?
Quote:

+ use of C provided performance and portability

AFAIR a large portion of the o.s. was written in assembly.
Quote:

+ modularity and code re-use rarely achieved today

Modern o.ses have shared objects for this, which are extensively used. Or do you mean something different?
Quote:

+ high and low level APIs

Forbid, Permit, Disable, Enable? Or (even much worse) macros to directly emulate those APIs? That was part (but not the only things) of the (very) bad design decisions which crippled the o.s..

Low-level APIs aren't necessary evil. But they made big mistakes here, unfortunately.
Quote:

+/- performance and responsive due to user level libraries, pointer based message passing, efficient interrupt handling and DMA, etc.

Pointer-based messages are part of the very bad design decisions.

However hooking the system interrupts and handlers is another very bad thing.

Performance and responsiveness comes from these decisions, plus the big pile of assembly code which was used in those critical parts of the o.s..
Quote:

- not enough abstraction and separation of functionality in some cases

Well, not only missing abstraction: opening up the doors to almost everything (including o.s.' structures) was the evilest thing ever...
Quote:

- not enough forethought into later/optional memory protection, resource tracking and security

That comes from a direct consequence of the very bad decisions which I was talking about.
Quote:

- AmigaDOS was inconsistent with the rest of the AmigaOS

Plus... BPTRs and BSTRs: how pervert can be an o.s. designer which introduced that completely dumb thing? Bah...
Quote:

The AmigaOS has aged better and with fewer changes than DOS with Windows, MacOS, Unix, RISC OS, Atari TOS with GEM, etc. Most of the operating systems of similar age were either replaced outright and/or mostly rewritten over time to become what they are today.

The Amiga didn't really aged: it's the same young guy. With all adolescent problems...

The other o.ses were reimplemented and evolved. But that's not a bad thing "per se", if you keep a good APIs compatibility.

Some, like MacOS, completely changed both interface (APIs) and implementation.
Quote:

The AmigaOS has not evolved as much because of an extreme focus on compatibility and lack of investment.

The reality is that the Amiga o.s. was completely bounded to its bad designs, which prevented an evolution. It should have made a process similar to what Apple did with MacOS, passing to OS X: completely rebuild it.

The chase for compatibility was a consequence of this, plus the lack of investments.
Quote:

Many of the negatives above have been partially dealt with but some changes are a tradeoff which will reduce performance and compatibility with minimal benefits.

Well, all the very bad things are still there, and you cannot change them unless you want to substantially write a new o.s..

Yes, the Amiga o.s. had very good performance, but it was achieved due to the foolish design.

@cdimauro

Quote:

Anyway, look at AROS: how many resources had the developers to build it from scratch and port to x86, x64, PowerPC, 68K, and ARM (works only from host now, but at least is running)?

We are talking about the NT route, not staying with the classic OS. AROS brings nothing in this regard.

@cdimauro

Quote:

Frankly speaking I prefer to pass as many parameters as possible in registers. The altenatives are: - pushing them to the stack; - fill a structure and pass it.

So would I prefer this. Stacking is such a C thing. Which means Intuition did it a lot even though leaving them in registers would have been more efficient. It also supported static structures. Until these were replaced by dynamic tag lists. Which are stored on the stack.

Quote:

Who cares? 80286 in protected was already quite different compared to 8086. 80386 was FAR AWAY from the 8086, even in the so called real mode ("equivalent" to how 8086 worked).

80286 was 16-bit. 80386 was 32-bit. But it was only at 80486 that it got basic big endian support. It was at that point and beyond that Intel really took off. When the 80586 was due it looked like Mototola had given up. Up to then Motorola CPUs were good and better in some respects.

To make a comparison. I see it like replacing the 68000, with a 65816, because it was also 16-bit. Although the 65816 never quite made it to the same standard.

I also read things on the internet claiming that Apple should have used Intel from the start. Which I think sounds stupid. The 80286 is said be from 1982 and 80386 from 1985. Using a 16-bit CPU instead of a modern 32-bit internal one because it had an Intel logo on it sounds even more stupid to me. But Apple started off with little endian so what are these people complaining about?

Quote:

Would you have used the real mode for porting any o.s., working exactly like 8086, or would you have written it to take advantage of all changes which the 80386 (and successors) introduced?

No way! But I also wouldn't gone for anything below an 80386. The time of Doom.

Quote:

Absolutely no. We were talking about porting an o.s. to another platform, even completely different from the original one.

The difference here is that Apple moved beyond Classic MacOS. On PPC is where they replaced MacOS with OSX AKA OSX86. Where as AmigaOS4 is still like Classic MacOS on PPC. We haven't got an Amiga OSX4 yet. So to speak.

Quote:

Therefore the RISC "philosophy", which worked quite well with few transistors and costly memories, wasn't superior anymore. Whereas the CISC designs, which allowed better code density (which positively impacts performances at all levels) and more "useful work" to be computed with a single instruction, gained more advantage.

In the end, it looked like x86 became a hybrid, with an external classic CISC opcode interface, but with RISC internals. The best of both worlds.

Quote:

OS4 NEEDS backwards compatibility because the port was never completed, and it still uses/runs some 68K code.

There are some other issues like licensing. For example they are restricted to ARexx in 68K only. Some other things like XAD from a third party were yet to be ported natively and bugs fixed. Last time I checked.

But, 68K compatibility was also a feature in OS4, a major feature. It also NEEDED to be 68K compatible to be relevant to the Amiga at all. Otherwise, it would have had nothing in common with it had they sugar coated the Amiga label sitting on top of a foreign OS by slapping Amiga on it.

There is already enough people to this day that ask what has the AmigaOne and OS4 got to do with the Amiga? And then go on about how it has nothing to do with the Amiga hardware. Well if you have to ask...

Quote:

On the contrary: Intel (and Microsoft) NEEDED to stay backward compatible due to the very large software base. There are still PCs which run DOS applications, either directly or using a virtualization layer or a virtual machine.

They should consider porting DOS to Windows. Are they stuck in the 80's? LOL. Well, emulation and now virtualisation exists. So, it can be done. Sure, using an emulator is a bit dirty, compared with running directly but it works. But hey, even the A1000 had DOS support!

I know this older gentleman who wrote a few BASIC programs in DOS. He would like to run them on his modern PC. Visual BASIC is worlds apart. So they break. I suggested running an old PC emulator but he doesn't seem to like that suggestion. I can see how stupid it looks running a PC emulator on a PC, like running UAE on an Amiga; but, the OS and CPU is worlds apart so it might be the best to do. Unless there is a newer BASIC than accepts old BASIC I haven't heard of. Perhaps similar thoughts kept x86 from being totally replaced. FUD.

Quote:

They had/have the license to port it to any architecture.

What's the meaning of this!?

https://amigaworld.net//modules/newbb/viewtopic.php?topic_id=19850&forum=17&start=0&viewmode=flat&order=0#311661

Finally, I found a reference to this. Been looking for so long!

Quote:

I was talking about the machines, so the hardware: hence there are no bitplanes supported in hardware (unless you like EGA and most VGA graphic modes).

It must have been lost in translation. When people think of VGA they tend to just think of chunky. In fact I've almost heard the AmigaOS on x86 argument as much as I've heard the AGA should have been VGA argument. The first understandable but the second shows a clear lack of Amiga hardware knowledge.

Quote:

Such video cards came from PCs...

How bizarre. And obviously there were chipsets built before they became the sole domain of PC video cards.

Quote:

Well, they aren't really little-endian, since they have only byte-accessible registers.

The hardware is little endian, as for example, a timer is in lo byte, hi byte order. But because of the way it is memory mapped all the registers must be accessed by the byte. So a 16-bit timer value needs to be split into "6502" order. Little endian tends to place emphasis on going byte-wise and building up from there.

Quote:

That was the point: a never finished port. Otherwise following the AROS approach you can have old 68K applications running on a (UAE-based) sandbox.

As pointed out above there are other reasons than it simply being incomplete. I hope on AROS it's improved as I tried the AROS 68K approach before and the most I got was some error about something missing. Unless it works out of the box there's no contest.

Quote:

Well, I guess that you never (low-level) programmed a Commodore 64:

No. I never owned a Commodore C64. But I did low level program my Commodore C16.

Quote:

Being fanatic is religious.

I'm religious about the Amiga.

Quote:

https://www.youtube.com/watch?v=eWG-nHuuCRc

Classic Barbarian. My god is better than your god discussion.

Quote:

Me too. Until I seriously technically analyzed all facts.

I think the Amiga held it for a long time. Even now I still think the copper deserves respect for being able to modify the screen as the raster draws it. A GPU blitting or compositing it all together in a back buffer just doesn't look as impressive. But it's obvious the Amiga would be eclipsed with development coming to a halt and the PC moving ahead. There was talk of catch up and then there was talk of running rings. Even now I am sitting here typing this on a laptop. A ten year old PC laptop. Dual 1.8 Ghz 64-bit like my X1000. But even FireFox works better on it. My X1000 used to feel fast but even now the bloated web sites make it slightly sluggish.

Quote:

Again: AROS.

And on that point. AROS was able to to throw off the shackles as it is a clone of AmigaOS that re-implements it. Unlike OS4 it doesn't work directly off any sources except that in the include files itself and anything else made public. So for its merits it can't go beyond anything official. If actual AmigaOS source code is considered official, which is why OS4 has that claim to fame. But this comes with good and bad points. OS4 sticks close to the API and has that responsibility burdened on it where as AROS only clones (or copies rather) the API so it's source compatible and portable. That allows AROS more freedom because a core value was not supporting legacy 68K programs directly but just naively supporting the hardware it is compiled for. Including x86 and beyond being an obvious target. Being source compatible as much as possible.

This also places OS4 in a hard position because it's tied to hardware that has become expensive to produce. But it's lacking in what it offers. It also seems to be in judgement from a 68K crowd but the point was to move beyond that. But it has yet reached the limit of that hardware. And now the hardware, even the X1000 has gone beyond what level OS4 is at.

It's obvious OS4 needs to move beyond. But in doing so, what would it become? It's easy to look at AROS, but, AmigaOS would need to be more than a clone of it's former self. Otherwise, what would be the difference between it and a copy?

Quote:

Because the Amiga o.s. was bad by design. I've already talked about it the past here. If you make a search you'll find my posts with much better (and many) technical facts about the argument.[/quote I think it was fine for the time. But I did see problems with the Workbench being messy and other things I don't recall which OS4 inherits and doesn't fix. But I hate searching on here, even using Google to work around searching bugs and limitations, as I rarely find what I'm looking for. [quote]P.S. Sorry but the quote tool didn't worked for the first message, so I had to manually copy & paste your writings, losing all smiles (and a lot time. Damn!). The forum is ageing, badly, like the Amiga o.s....

I tend to copy the smileys by hand. But a backup I use is the Quote button I can see below. But I find IBrowse, despite being archaic, is still the best. Unlike Odyssey it doesn't place major limitations on the Quote box. And I can go back then forwards and the post is still there. Odyssey, despite being more modern, is old fashioned by comparison. Limited Quote length, I made a thread about the problem, and if I go back and forward again it deletes the text! FireFox doesn't seem much better. There I am typing when suddenly it loads up some ad for Linux Mint. If I can go back, though most times the link is lost, it deletes all I typed! Argh. If it wasn't for the TextCache plugin I'd be stuffed.

@Hypex

Quote:

I also read things on the internet claiming that Apple should have used Intel from the start. Which I think sounds stupid. The 80286 is said be from 1982 and 80386 from 1985. Using a 16-bit CPU instead of a modern 32-bit internal one because it had an Intel logo on it sounds even more stupid to me. But Apple started off with little endian so what are these people complaining about?

I think maybe these people are talking about the later incarnation of Apple, meaning they should have dropped PowerPC for Intel once it became clear the next Apple OS would be based on NeXTSTEP (which was already on x86).

Regarding ARexx and OS4... can't Hyperion just license it from IBM?
(Hahaha... hehe.. uhm.)

Last edited by kolla on 04-Aug-2018 at 02:36 PM.

_________________
B5D6A1D019D5D45BCC56F4782AC220D8B3E2A6CC

@kolla

Quote:

I think maybe these people are talking about the later incarnation of Apple, meaning they should have dropped PowerPC for Intel once it became clear the next Apple OS would be based on NeXTSTEP (which was already on x86).

I don't have a source to cite but from what I recall by the context it was referring to the Mac from the outset. It was written by someone who was a technology writer but obviously didn't understand the hardware at a technical level.

Quote:

Regarding ARexx and OS4... can't Hyperion just license it from IBM? (Hahaha... hehe.. uhm.)

Well yeah they could if it still can be. But it would cost them I bet.

I like ARexx. It is a fairly understandable and dynamic language. Also, ARexx ports offer features for free. Just need to pass a data structure with command strings to an application and it can process it without using ARexx. Just interpret a string.

@pavlor

Quote:

pavlor wrote: @cdimauro Quote: Anyway, look at AROS: how many resources had the developers to build it from scratch and port to x86, x64, PowerPC, 68K, and ARM (works only from host now, but at least is running)? We are talking about the NT route,

Not only that. And NT was already quite usable from v3.5. It missed DirectX and games support, but Win32 applications ran fine (for the reasons that I've also explained in my posts which I was talking about).
Quote:

not staying with the classic OS. AROS brings nothing in this regard.

Here's your statement:

"With such power, even I would be able to port AmigaOS to x64 in two years time. "

Porting the Amiga o.s. to x64 is NOT a gargantuan task and, in fact, it was made by the AROS team.

Let me remark it: PORTING.

And the AROS team did MUCH more: they didn't ported the Amiga o.s., but completely rewritten it from scratch.

P.S. Again, this has nothing to do with Windows NT.

@ thread

Why was AmigaOS 4.X developed only for PowerPC?

Because they knew POWER9 would be along someday.

_________________
Tinkering with computers.

@Hypex

Quote:

I like ARexx. It is a fairly understandable and dynamic language.

ARexx is fine, but only because we don't have APython. Or even AAwk.

The Rexx format command, for example, is interesting, and even innovative in a way, but I'd rather use printf. The indication of a language becoming mainstream is when the designers finally give in and add printf.

Last edited by bison on 04-Aug-2018 at 04:59 PM.

_________________
"Unix is supposed to fix that." -- Jay Miner

@Hypex

Quote:

Hypex wrote: @cdimauro Quote: Frankly speaking I prefer to pass as many parameters as possible in registers. The altenatives are: - pushing them to the stack; - fill a structure and pass it. So would I prefer this. Stacking is such a C thing. Which means Intuition did it a lot even though leaving them in registers would have been more efficient. It also supported static structures. Until these were replaced by dynamic tag lists. Which are stored on the stack.

If you have complex structures ("objects") then a dynamic list is a better general, and future-proof, mechanism.

Another alternative is having a fixed structure (and/or a set of fixed parameters), and then getters and setters for extra parameters/attributes.
Quote:

Quote: Who cares? 80286 in protected was already quite different compared to 8086. 80386 was FAR AWAY from the 8086, even in the so called real mode ("equivalent" to how 8086 worked). 80286 was 16-bit. 80386 was 32-bit.

The 80286 had a 16-bit ALU, but 32-bit pointers for the largest memory model (with 1GB of virtual address space, and 24 bit of physical memory addressable).

The 80386 had 32-bit and 48-bit respectively, but the most common memory model was a flatten, 4GB, one.
Quote:

But it was only at 80486 that it got basic big endian support. It was at that point and beyond that Intel really took off. When the 80586 was due it looked like Mototola had given up. Up to then Motorola CPUs were good and better in some respects. To make a comparison. I see it like replacing the 68000, with a 65816, because it was also 16-bit. Although the 65816 never quite made it to the same standard.

Hum. Questionable. The 80386 did so well (even without the BSWAP instruction) that Intel decided to "kill" the agreements/contracts with third-party vendors like AMD (used to be second suppliers for IBM, for example), which then had to reverse engineer and recreate from scratch the missing microcode (Intel gave them up to the 80286 one).

Regarding the comparison with Motorola, the main problem is that the 68000 successors came always late compared to the "equivalent" Intel product. And starting from the 80486, Intel released versions with much higher clocks, which compensated and outperformed the Motorola processors at the end.
Quote:

I also read things on the internet claiming that Apple should have used Intel from the start. Which I think sounds stupid. The 80286 is said be from 1982 and 80386 from 1985. Using a 16-bit CPU instead of a modern 32-bit internal one because it had an Intel logo on it sounds even more stupid to me. But Apple started off with little endian so what are these people complaining about?

I think that it's a fake. I don't give a penny to this theory.
Quote:

Quote: Therefore the RISC "philosophy", which worked quite well with few transistors and costly memories, wasn't superior anymore. Whereas the CISC designs, which allowed better code density (which positively impacts performances at all levels) and more "useful work" to be computed with a single instruction, gained more advantage. In the end, it looked like x86 became a hybrid, with an external classic CISC opcode interface, but with RISC internals. The best of both worlds.

With a big difference: the RISC is internal, and you cannot access it. It's an implementation detail (albeit important).

What's exposed outside is the ISA, and that's the most important thing: the one which matters.
Quote:

Quote: They had/have the license to port it to any architecture. What's the meaning of this!? https://amigaworld.net//modules/newbb/viewtopic.php?topic_id=19850&forum=17&start=0&viewmode=flat&order=0#311661 Finally, I found a reference to this. Been looking for so long!

I think that this is related to the code which was developed by Hyperion contractors. Hyperion has only PowerPC licenses about that work.

But AFAIR the agreement between Amiga Inc and Hyperion is not restricted to PowerPC, and Hyperion should have the right to use the Amiga o.s. 3.1 sources for whatever they want to do, included porting to x86 or x64. But in this case either they have to license the new code from its contractors (due to the above PowerPC license) or completely rewrite it.
Quote:

Quote: That was the point: a never finished port. Otherwise following the AROS approach you can have old 68K applications running on a (UAE-based) sandbox. As pointed out above there are other reasons than it simply being incomplete. I hope on AROS it's improved as I tried the AROS 68K approach before and the most I got was some error about something missing. Unless it works out of the box there's no contest.

I never tried Amibridge/Janus, so I cannot answer here.
Quote:

Quote: Well, I guess that you never (low-level) programmed a Commodore 64: No. I never owned a Commodore C64. But I did low level program my Commodore C16.

The processor is very similar, so you should know its very complex instructions.
Quote:

Quote: Being fanatic is religious. I'm religious about the Amiga.

I'm atheist. About everything.
Quote:

Quote: Again: AROS. And on that point. AROS was able to to throw off the shackles as it is a clone of AmigaOS that re-implements it. Unlike OS4 it doesn't work directly off any sources except that in the include files itself and anything else made public. So for its merits it can't go beyond anything official. If actual AmigaOS source code is considered official, which is why OS4 has that claim to fame. But this comes with good and bad points. OS4 sticks close to the API and has that responsibility burdened on it where as AROS only clones (or copies rather) the API so it's source compatible and portable. That allows AROS more freedom because a core value was not supporting legacy 68K programs directly but just naively supporting the hardware it is compiled for. Including x86 and beyond being an obvious target. Being source compatible as much as possible.

AROS is 68K binary compatible, and you can have also an AROS' Kickstart ROM (albeit it takes more than 512KB).

BTW, AROS is not an Amiga o.s. clone. The goal is to have (at least) v3.1 compatibility, but new things were also introduced.
Quote:

This also places OS4 in a hard position because it's tied to hardware that has become expensive to produce. But it's lacking in what it offers. It also seems to be in judgement from a 68K crowd but the point was to move beyond that. But it has yet reached the limit of that hardware. And now the hardware, even the X1000 has gone beyond what level OS4 is at. It's obvious OS4 needs to move beyond. But in doing so, what would it become? It's easy to look at AROS, but, AmigaOS would need to be more than a clone of it's former self. Otherwise, what would be the difference between it and a copy?

They are already different because OS4 diverged from the 3.1 roots. The same happens to MorphOS and AROS, but at the least they are more 3.1 compatible from a developer point-of-view.

Anyway, OS4 has no future. Its development is stalled from many years, and I hardly doubt that the "Duke Nukem Forever" version (AKA 4.2) will ever bring all promises that Hyperion made.

MorphOS seems to be still healthy. But I don't know how much they can continue sticking to the PowerPCs, which are dead (and not even from now).

AROS lacks developers, but it's open sources and there is at least the chance that someone can take them and contribute. In fact, AROS-Exec is still active.
Quote:

Quote: P.S. Sorry but the quote tool didn't worked for the first message, so I had to manually copy & paste your writings, losing all smiles (and a lot time. Damn!). The forum is ageing, badly, like the Amiga o.s.... I tend to copy the smileys by hand. But a backup I use is the Quote button I can see below. But I find IBrowse, despite being archaic, is still the best. Unlike Odyssey it doesn't place major limitations on the Quote box. And I can go back then forwards and the post is still there. Odyssey, despite being more modern, is old fashioned by comparison. Limited Quote length, I made a thread about the problem, and if I go back and forward again it deletes the text! FireFox doesn't seem much better. There I am typing when suddenly it loads up some ad for Linux Mint. If I can go back, though most times the link is lost, it deletes all I typed! Argh. If it wasn't for the TextCache plugin I'd be stuffed.

It's not a browser problem, but an issue with the Amigaworld forum. Sometimes happens that if you click on the "quote" button (below the edit box), it didn't copy the text from the original message. And manually copying the smiles is a no-go for me...

@Signal

Quote:

Signal wrote: @ thread Why was AmigaOS 4.X developed only for PowerPC? Because they knew POWER9 would be along someday.

Then start porting to it. And customers should prepare big buckets for buying a POWER 9 platform.

However POWER 9 hasn't very good single core/thread performances. And OS4 uses just one processor core/hardware thread...

@bison

Quote:

bison wrote: @Hypex Quote: I like ARexx. It is a fairly understandable and dynamic language. ARexx is fine, but only because we don't have APython. Or even AAwk. The Rexx format command, for example, is interesting, and even innovative in a way, but I'd rather use printf. The indication of a language becoming mainstream is when the designers finally give in and add printf.

Well, Python offers not only a printf-like equivalent. That's why is THE "glue" language to look at.

Python is the swiss-army of nowadays, whereas AREXX lost its time...

@bison

OS4 has a Python port. IIRC it's included in the OS. Looks like they are trying to supplant ARexx with it. But a printf function in non-C? Oh no.

@cdimauro

Quote:

Porting the Amiga o.s. to x64 is NOT a gargantuan task and, in fact, it was made by the AROS team.

Next time, I should be more precise... I meant porting with full OS4.x compatibility and all modern features (SMP, memory protection etc.).

Quote:

And the AROS team did MUCH more: they didn't ported the Amiga o.s., but completely rewritten it from scratch.

And the outcome? Same old classic OS with the very same limitations as the original one.

Quote:

P.S. Again, this has nothing to do with Windows NT.

Again, my point is very much about WinNT: (nearly) full backwards compatibility with all modern features. AROS shares the same old heritage (limitations) as other Amiga-like OSs, without full OS3.x software compatibility (except quite obscure 68k version). Only strong side of AROS is support for subset of mainstream hardware.

We don´t need another AROS (eg. MorphOS on x64), what we really need is AmigaOS NT and there are no resources to complete such goal.

@pavlor

Quote:

pavlor wrote: Next time, I should be more precise... I meant porting with full OS4.x compatibility and all modern features (SMP, memory protection etc.).

Full OS4.x compatibility AND all modern features are NOT possible at all: you have to chose either the first (but losing the 68K applications integration. This is only possible with a 32-bit big-endian architecture, which clearly x64 isn't) or the second one (in this case it's not a "port", but a brand new o.s.).
Quote:

And the outcome? Same old classic OS with the very same limitations as the original one.

Not all the same: AROS x64 is capable of addressing A LOT of memory, and has some experimental SMP support.
Quote:

Again, my point is very much about WinNT: (nearly) full backwards compatibility with all modern features. AROS shares the same old heritage (limitations) as other Amiga-like OSs, without full OS3.x software compatibility (except quite obscure 68k version). Only strong side of AROS is support for subset of mainstream hardware. We don´t need another AROS (eg. MorphOS on x64), what we really need is AmigaOS NT and there are no resources to complete such goal.

As said before, you're NOT asking a port, but a new o.s. (inspired by the Amiga o.s., sure, but still absolutely NOT the same).

The comparison with Windows NT is NOT possible at all, because Windows had, since the first version (believe me or not, this is the reality. I've talked quite extensively about it in my old posts, providing links to technical sources that proved it), the necessary abstraction at the o.s. and application level, that allowed to introduced the modern features keeping the essentially the same compatibility at the API level.

As I said many times, this is NOT possible with the Amiga o.s. (and successors/reimplementations) due to the very bad design decision which were made by the engineers which created it, and which completely crippled the o.s. and its evolution.

TL;DR: a Amiga o.s. port to x64 is possible and AROS clearly has proven it. Introducing modern o.ses features is not, due to bad design decisions. For more, in depth, information, please take a look at my old messages.

@cdimauro

Quote:

Full OS4.x compatibility AND all modern features are NOT possible at all: you have to chose either the first (but losing the 68K applications integration. This is only possible with a 32-bit big-endian architecture, which clearly x64 isn't) or the second one (in this case it's not a "port", but a brand new o.s.).

Give me the same resources MS had for NT and I prove I´m right.

Take new kernel, add well integrated sandbox here and there and provide the same UI as OS4... and you have perfect AmigaOS with (nearly) full compatibility and modern features. We don´t need the same level of 68k/PPC integration as OS4, isolated enviroments sharing the similarly looking UI would be more than enough. Eg. OS4/OS3 applications run from the NT WB, can use - from the user´s point of view - the same public screens as the NT applications and their GUI looks like GUI of NT applications, but are entirely isolated from the NT world.

Well, I could realize the above idea with far more modest budget than MS.

@pavlor

Quote:

pavlor wrote: Give me the same resources MS had for NT and I prove I´m right. Take new kernel, add well integrated sandbox here and there and provide the same UI as OS4... and you have perfect AmigaOS with (nearly) full compatibility and modern features.

No, you don't, you can't, even if you get the current Microsoft capital value.

You asked for "full OS4.x compatibility and all modern features" (your words) and this is simply IMPOSSIBLE. Because there are several TECHNICAL issues.

Even if you try lo lower the bar with "(nearly) full compatibility" it's still NOT possible.

The problem is the Amiga o.s. design itself, as I've already talked about.

If you want modern features you need a NEW o.s. which works VERY DIFFERENTLY from the Amiga one. Hence: you lose backwards compatibility.

That's something which an experienced Amiga developer knows, but if you don't trust me then you can ask some Amiga o.s. 3.1, OS4, MorphOS, AROS core developer. Some are users of this forum (or even on other forums).
Quote:

We don´t need the same level of 68k/PPC integration as OS4,

This is the first thing that OS4 and MorphOS will loose by moving to a little-endian ISA and/or a 64-bit one.
Quote:

isolated enviroments sharing the similarly looking UI would be more than enough. Eg. OS4/OS3 applications run from the NT WB, can use - from the user´s point of view - the same public screens as the NT applications and their GUI looks like GUI of NT applications, but are entirely isolated from the NT world. Well, I could realize the above idea with far more modest budget than MS.

Me too, and that's something which I'm thinking about from some years now, but I had no time to work on it (not so much free time, and the free time is already fully booked with other projects).

Anyway what you describe is completely different from your original statement.

Of course, if you change the requirements then the effort is GREATLY reduced, but... you're changing the cards on the table while playing. Not fair...

@cdimauro

Quote:

Anyway what you describe is completely different from your original statement.

I don´t think so. If it looks like AmigaOS, runs Amiga applications like AmigaOS and has general "feeling" of AmigaOS, then it is AmigaOS by my definition.

Quote:

you're changing the cards on the table while playing. Not fair...

You know, I never played your game. Users don´t care about "things under the hood", they want to run their software in familiar enviroment. You may call this cheating, for me it is sensible progress.