Submit your best idea for the most realistic path to open source Amiga OS 3.x

[OlafSch]

@eXeler0

I am not on amigamap and certainly lots of other peoples are not either

and now? What do we do?

You claim 200, I say much more, my word against yours

and now?

discussion is really senseless... you have not contributed anything, you are not interested in Aros and never will, just stop spreading wrong informations about Aros when you do not know (the same is true for others here) or downtalking efforts of others who really do something. I wish you good luck with open sourcing AmigaOS and will take off my (not existing) hat if you really manages to...

[Zetr0]

Cat amongst the pidgins idea -

Why not port 3.x and or 4.x to x86?

Money would be made - shits would be given and possibly a very cool x86 would produce some wonderful creativity!

[Minuous]

@OlafSch:

Well, if you hate AmigaOS that much and want to strike out in an incompatible direction, good luck against all the other free x86/x64 OSes in this day and age when it is limited to OS3.1+AROS-only extensions.

I don't see where the userbase and new development tools would come from; Amiga users are likely to continue with OS3.9/4.1/MOS and UNIX users would see no advantage in switching to AROS either.

[wawa]

Originally Posted by Minuous

MCE and Report+, there are probably others also. Limiting applications to using only ancient ClassAct features instead of making use of improvements made in OS3.5/3.9/4 is not the way forward. The point is: apps wanting to be compatible with AROS have to be crippled to not use all AmigaOS features; an OS replacement should be able to replace that OS and run all its programs. Just replacing OS3.1 with something bigger and slower is probably not very useful to most people, especially now that the real OS3.1 source code has leaked.

And the situation with AROS x86 is even worse as it doesn't support emulating any 68K/PPC code.

just checked reportplus from aminet with latest aros68k nightly and classact. it works. wanna screenshot?? :

https://1drv.ms/i/s!AuvzEgbTrkxP-F8NxsNKO7zx3XDy

your point has been proven wrong. and also it proves that you are spreading your wrong preassumptions as facts without even checking. this is fud and bad will. period.

[OlafSch]

Originally Posted by Minuous

@OlafSch:

Well, if you hate AmigaOS that much and want to strike out in an incompatible direction, good luck against all the other free x86/x64 OSes in this day and age when it is limited to OS3.1+AROS-only extensions.

I don't see where the userbase and new development tools would come from; Amiga users are likely to continue with OS3.9/4.1/MOS and UNIX users would see no advantage in switching to AROS either.

Many people (expecially when using real hardware) do not use and like 3.9

MOS will make a ISA switch either with radical changes and then propably be even less compatible than AROS. And for some strange reasons 68k software partly runs on both Aros (68k) and MOS but not 4.1... weird world. Could it be that they also changed a lot of things and dropped compatibility... questions over questions

But it is not important because it is more propable that hell freezes than you using Aros but (as I wrote before) stop spreading nonsense and talking about things you have no clue of. Aros obviously belongs to that.

[Minuous]

>more propable that hell freezes than you using Aros

If it remains at OS3.1 levels of functionality of course I won't use it, why would anyone!? If the goal is OS3.9 and there are problems with reaching that, I would understand, I just don't get what is so magical about OS3.1 that once OS3.1 functionality had been achieved there was no interest in supporting later and better AmigaOS versions. It is not like Windows where every update is worse than the one before :-/

>And for some strange reasons 68k software partly runs on both Aros (68k) and MOS but not 4.1... weird world. Could it be that they also changed a lot of things and dropped compatibility... questions over questions

ReAction-based programs can easily be built for OS3.9, OS4.1 and MOS but not AROS x86. The changes in OS4.1 are mostly different APIs to access the same functionality. With AROS x86 the functionality is just missing.

>things you have no clue of.

I've installed and run it several times but as it is incomplete and has hardly any software for it I didn't find it of use and don't think it represents the most viable path for the Amiga going forwards. There is no need to flame me for just pointing that out; I don't hold you or any single person responsible.

[wawa]

Originally Posted by eXeler0

I answered that in the post directed at you. Its about the convenience of building on something already present in so many users Amigas..

in oreder to improve "what is present in so many peoples amigas" you would have to recompile it first. same as aros. then you would have to find a way to deliver replacement roms or make your system softkick on old ones (only from drives accessible by the genuine kicksart). same as aros..

you want to build on 3.x foundation no matter what, for sentiment reasons. then say it. dont try to invent technical execuses.

If no one wants to merge, then fine. The 3.x branch would stand on its own feet just fine.

you will have to refraiun to gray zone of half-legality and asm patching paricular binaries and libs by hand to gain some 1% speedup as it is already done now on this very forum among others.

its surely not on us who wont cooperate.

I'm not exactly sure how your thinking goes..

exactly like you say, you estimate the numbers of users upon some amiga map. which proves procentually all amigalike-ng flavours suffer the same low frequency.

genuine amiga will always be most numerous. maybe vampire will become more popular than the others. i hope so.

[wawa]

Originally Posted by Minuous

I've installed and run it several times but as it is incomplete and has hardly any software for it I didn't find it of use and don't think it represents the most viable path for the Amiga going forwards.

apparently you have not checked for more than five minutes, if you state that something that works doesnt work. if it didnt most likely the stack was too low.

now. if you dont think aros is a viable way forward, you are welcome to tell us what you think is.

[OlafSch]

Originally Posted by Minuous

>more propable that hell freezes than you using Aros

If it remains at OS3.1 levels of functionality of course I won't use it, why would anyone!? If the goal is OS3.9 and there are problems with reaching that, I would understand, I just don't get what is so magical about OS3.1 that once OS3.1 functionality had been achieved there was no interest in supporting later and better AmigaOS versions. It is not like Windows where every update is worse than the one before :-/

>And for some strange reasons 68k software partly runs on both Aros (68k) and MOS but not 4.1... weird world. Could it be that they also changed a lot of things and dropped compatibility... questions over questions

I find many programs will run on OS3.9, OS4.1 and MOS but not AROS x86. The changes in OS4.1 are mostly different APIs to access the same functionality. With AROS x86 the functionality is just missing.

>things you have no clue of.

I've installed and run it several times but as it is incomplete and has hardly any software for it I didn't find it of use and don't think it represents the most viable path for the Amiga going forwards. There is no need to flame me for just pointing that out; I don't hold you or any single person personally responsible.

Strange... I have hundreds of programs for it including word processors like Final Writer, Raytracers and so on. Most compatible programmed software runs on it without problems. I assume you talk about X86 now, we here discussed about 68k. And I know there pretty good what is working and what is not working... Most 68k software works on X86 either, just that you have to use UAE for it. But that is the price to pay if you want something improved. MOS devs have announced too that they will use UAE for 68k. Solutions like Petunia simply would require too much work.

I think you guys do not understand it or do not want to understand it... at least on 68k it feels for software like running in 3.1 so you do NOT need special aros software on it. Thus there is plenty of software available...

regarding amiga users... facebook group "commodore amiga" has more than 11.000 users, much more than on amigamap. If that are users depends on definition, I assume majority uses emulation

[wawa]

Originally Posted by Zetr0

Cat amongst the pidgins idea -

not yet again!!

Why not port 3.x and or 4.x to x86?

because the owners or licesnees do not wan t it!
besides it has been done. practically. its aros.. damn

[wawa]

Originally Posted by Minuous

good luck against all the other free x86/x64 OSes

this doesnt make any difference. neither aros, nor os4 nor morphos has any chance in comparison. no matter on what hardware. still aros is better off even here, as it is portable.

[OlafSch]

@wawa

lol

:d

[Minuous]

Yes, AROS 68K will run most OS3.1 apps, and also most OS3.5/3.9 apps if the relevant missing bits from AmigaOS 3.9 are copied over.

But the thread is about how to open source AmigaOS. OS3.1 has already been leaked, so OS3.5 and later are what now needs to be open sourced. AROS intends to be limited to OS3.1 plus some AROS-only extensions. I don't see how that will help with open sourcing AmigaOS. The OS3.9/4 source has to be purchased from H&P/Hyperion, recreated by decompilation or rewritten by revising the OS3.1 sources (or from scratch in the case of new components).

[OlafSch]

@Wawa

:-)

[OlafSch]

Originally Posted by Minuous

Yes, AROS 68K will run most OS3.1 apps, and also most OS3.5/3.9 apps if the relevant missing bits from AmigaOS 3.9 are copied over.

But the thread is about how to open source AmigaOS. OS3.1 has already been leaked, so OS3.5 and later are what now needs to be open sourced. AROS intends to be limited to OS3.1 plus some AROS-only extensions. I don't see how that will help with open sourcing AmigaOS; like I said before the OS3.9/4 source has to be purchased from H&P/Hyperion, recreated by decompilation or rewritten from scratch.

3.1 is leaked but nobody can do something with it, at least not legal and in public

the sources are open in a way but it is not "open source"

3.5 and 3.9 is a mixture of H&P and third-party developments that later was licensed and integrated in 4.X. I do not know the contracts between Hyperion and those developers but I assume that Hyperion is not very intereted in it. And Mr. Haage (from my last contacts) is not "very" interested in amiga stuff anymore to say it politely

If you really think there is a chance go for it

you can contact Mr. Haage easily
http://www.haage-partner.de/

the third-party developers who contributed to 3.5 and 3.9 should be somewhere in documentation
a number of them even still active on the different platforms. Contact them and ask them if they would open source their software

[wawa]

Originally Posted by Minuous

if the relevant missing bits from AmigaOS 3.9 are copied over.

why are insisting on.. i am sorry to say, lying on a purpose to support you prejudice?

class act is not a part of os3.9. its a freely available third party contribution. same as third party mui classes you may be using with your os3 or aros respectively.

since aros is not an option here, i will now be off from this thread except if you further insist to spread wrong information about aros, others may take for truth.

good luck with that leaked source.

[eXeler0]

Originally Posted by OlafSch

@eXeler0

I am not on amigamap and certainly lots of other peoples are not either

and now? What do we do?

You claim 200, I say much more, my word against yours

and now?

discussion is really senseless... you have not contributed anything, you are not interested in Aros and never will, just stop spreading wrong informations about Aros when you do not know (the same is true for others here) or downtalking efforts of others who really do something. I wish you good luck with open sourcing AmigaOS and will take off my (not existing) hat if you really manages to...

I'll gladly accept the correct version of any wrong info I have spread about AROS.
Regarding the number of users, I already stated the source of info and added that it might be wrong.


Good luck with your AROS-centered life.

To thread: Next post better be back on topic :-)

[OlafSch]

Originally Posted by eXeler0

I'll gladly accept the correct version of any wrong info I have spread about AROS.
Regarding the number of users, I already stated the source of info and added that it might be wrong.


Good luck with your AROS-centered life.

my life is not AROS centered I must correct you... it is Windows-centered (regarding computer) and besides Computers are not life anyway...

in opposite to many others here (including you?) it is just a hobby not religion

talk is cheap... go make amiga os open source, find out who the correct copyright owner is, negotiate a price and then setup a bounty to free the sources. Then make a new thread here

I am out too... good luck with it

[matthey]

Originally Posted by Zetr0

Cat amongst the pidgins idea -

Why not port 3.x and or 4.x to x86?

Money would be made - shits would be given and possibly a very cool x86 would produce some wonderful creativity!

As Olaf hinted at, compatibility would be lost without sandboxing 68k programs (with an emulator like UAE) as AROS x86 did. Little endian (LE) would be a major porting hurdle to overcome. The x86 ISA only has 8 GP registers limiting Amiga like register usage or use x86_64 and have more problems with 64 bit pointers and code density deteriorates to RISC like. You are not going to have an efficient little AmigaOS in the end. If you go big, you need expensive high end hardware because smaller is faster and more energy efficient. As soon as an OS is on x86/x86_64 hardware everyone expects drivers for all the commodity hardware (some lacking info to program drivers) where the AmigaOS would not be competitive. I suppose SMP, memory protection and resource tracking support could be added at the same time as 64 bit and the LE change since all compatibility would be broken anyway. This would put the AmigaOS ahead of AROS x86 except that AROS x86 already has some software (mostly ports). IMO, you would have nothing competitive, nothing unique and nothing innovative. Many Amiga users would not even make the transition to such a platform.

Is it even necessary to go big with 64 bit physical addressing? Pointers need twice as much space and 8 byte alignment gives better performance with 64 bit pointers so more space is wasted aligning data (and code). The DCache is filled with many NOPs and zeros for only a few extra bits of addressing (usually only 40 bits of physical addressing is used). Code density drops significantly when moving to 64 bit wasting valuable ICache. Less than 10% of code and data is used often (probably much less on a small footprint OS like the 68k) so an MMU could quickly swap data between memory >4GB and the 32 bit 68k physical memory (memory<->disk is slow but memory<->memory is relatively fast) with the only limitation being that only 4GB of memory could be addressed at a time per task/process (this is a small limitation on an OS with a small footprint). The savings could be large for a multi-core CPU like an enhanced 68k where a 16kB L1 ICache and 16kB L1 DCache per core would likely significantly outperform that for most 64 bit ISA hardware (smaller caches are faster to access too). IMO, the 68k AmigaOS is a good candidate for such hardware as the AmigaOS does not currently use the MMU, backward compatibility could be maintained without sandboxing and the footprint of the AmigaOS is small enough that little MMU swapping would be necessary anyway and the 68k can improve code density to reduce L1 cache requirements and maximize the advantage of such a setup. An FPU and SIMD would still be able to process 64 bit datatypes. I don't want the 68k Amiga to follow the crowd to big and slow. I want it to be small, fast and friendly (lean and mean in a nice way).

Originally Posted by OlafSch

Many people (especially when using real hardware) do not use and like 3.9.

Most of the people who chose to stay with AmigaOS 3.1 have low end 68k hardware. The AmigaOS 3.9 glow icons are bigger, slower and don't look better with ECS/AGA unless slower screen modes are used (PeterK's icon.library partially offsets this disadvantage). More memory is required as some modules grew and a double reboot is necessary without a new kickstart ROM. The changes were good overall and a move in the right direction but Amiga users with high end hardware (RTG and accelerators) benefited more. It would be possible to optimize the 68k AmigaOS so that an AmigaOS 3.9 enhanced AmigaOS is faster and uses less memory than AmigaOS 3.1.

Originally Posted by wawa

class act is not a part of os3.9. its a freely available third party contribution. same as third party mui classes you may be using with your os3 or aros respectively.

Reaction is a part of AmigaOS 3.9 and is only a moderately enhanced ClassAct.

[wawa]

Reaction is a part of AmigaOS 3.9 and is only a moderately enhanced ClassAct.

i know that, but it isnt (yet) the os component by itself. it is freely avaliable free of charge. the point being, aros (or whatever alternative open source system you might want to create) doesnt need to recreate each and every amiga library, device or program if they are anyway available for the general public, except you need to improve or extend its functionality. what would be the sense of it otherwise? aros68k is done particularly to be compatible with all these bits ad pieces in mind. i think it is an advantage not a handicap, as minous try to make it look like. but lets leave this apparent offtopic be.