AROS 3D with Deadwoods latest abi v.0 build.

nikos · 2093

cdimauro

  • Junior Member
  • **
    • Posts: 50
    • Karma: +1/-1
Reply #45 on: December 27, 2020, 11:10:05 PM
(albeit, as I've already said, the specs aren't yet finalized).
Who writes the specs? Does he have coordinates?
If the answer is: everyone writes specifications, then this is the same answer: nobody writes specifications.
No specifications = No plan = No future = No ABIiV1

What was planned to be achieved had to be agreed and described BEFORE the development of ABIv1 began.
If the developers don't know this even now, after so many years, then what should users expect?
AFAIR the specs are written by the core developers, after discussions. They are already available, as Mazze reported, but not yet finalized.


The answer is very simple and very well known, and it was already reported tons of times: AROS developers work on what they like.Which is mostly ABIv1, because that's what they think is the future of AROS (albeit, as I've already said, the specs aren't yet finalized).ABIv0 stays and lives only because there is people like deadwood which is backporting what it's being made on v1. So, enjoy it until this work is being done.Currently the main problem for the ABIv1 is that developers aren't actively working on the most common and promising architectures (x64 AKA x86-64, ARMv7, ARMv8), but are involved on the less interesting, already died long time, and without a future: the 68K one.What's worse is that they forked AROS and are removing (if not even already done) the HIDD and rewriting parts of the o.s. to mimic the crappyness of the original Amiga o.s. (which is strictly tight to the chipset) just to improve the performance on this very under powered platform, which is struggling otherwise (and which is expected, since it's an ultra-outdated hardware: "Pentium" class).Years an years of work to abstract AROS from the hardware which are thrown into the trash can...
Cdimauro before posting you should have a clue what you talk aboutWhich of the Aros devs is involved in 68k currently? Kal not (you see it on changelog). Schulz is developing a kind of Amithlon-solution on ARM. Both are not involved in the vampire related fork. None of the devs involved there was or is active on the main branch, they are new.
It was written by Gunnar (AFAIK), in the Apollo forum, that Kal was contributing to the Vampire, and that two boards were sent to Kal and Schulz.

So, I was assuming that they were involved in the project and in the fork.
Quote
And they are only active because of 68k. Nothing is thrown in the trashcan, both are different forks.
A fork isn't a good thing, because it's a waste of resources.
Quote
It is planned to give back changes but those will certainly adding features or removing bugs.
Changes should be given back as soon as a binary is released.
Quote
The structure of Aros will not change.
Then what happens to the HIDD in the fork?
Quote
And if 68k or other architctures will be more successful has to be seen.
68K (as well as PowerPC) is a dead architecture, so there cannot be a discussion about which architecture is successful: certainly not the 68K one.

When someone with the knowledge as Deadwood have, and you question what is right or wrong, you better come up with some dammed good arguments why he stick with ABI v.0!! and do not want to get involved with the ABI v.1 versionI'm not a developer, but ABI v.1 is for sure a mess!!For political reasons he does not want to make a statement about that. It is not war!! if you do not like abi v.0 stay away from it! It is an open source platform, and people do as they want with it.I followed this project almost since the beginning, and one thing I learned is that what might be promised, might never happen!We are not going to take over the world anyway! It is a fun platform that is in the spirit of AmigaOS, and the most powerful there will ever be! Forget multi core, SMP, memory protection, bla bla. It is not going to happen. If that is important to you f... off. End of story!
SMP or 64bit needs dedicated software, who will write it?
Not necessarily. There's a reason why a pillar of AROS is source-level compatibility (binary-compatibility is only about 68K).
Quote
MP breaks almost any amiga software.
True. This (as well as resource tracking, and other things) cannot be changed.
Quote
You would start at zero then, a modern platform but without software except some 68k software running in emulation. That would hardly convince new users.
Nobody is talking about writing a new o.s. here: the AROS's goal is different.
Quote
As I wrote... amiga users must feel at home that means a environment including 68k that makes people forget that they are sitting at a PC and not a new amiga.
paolone already replied to you about it: Amibridge is what you need for integrating 68K stuff in a trasparent way.


And AROS is already giving you the same Amiga o.s. feeling...



OlafS3

  • Junior Member
  • **
    • Posts: 63
    • Karma: +3/-0
Reply #46 on: December 28, 2020, 12:06:09 AM
@cdimauro

That is exactly what I meant with "no clue". That was the case end of 2019 and from this time the post of Gunnar. Kal left all related to vampire if I remember right December 2019. Yes he got a board. The same is true for MSchulz. He got a board and indeed contributed some time but then also left. He has his own project now (a kind of Amithlon on ARM based on Aros 68k). Then Aros was forked, the dev contributing now are involved because they are interested in Aros 68k on Vampire, nothing else. Kal on the other hand has different views so, in my view, forking was best solution. The forked branch does not affect the main branch, even if something is removed there. And regarding future, of course anything 68k related cannot become mainstream again but in amiga terms the 68k market is multiple of all NG related platforms together regarding sales and number of users. How many active Aros (X86) users are there currently should be known to you. BTW I personal do not see the real potential for another exotic OS out there. It is difficult today to really attract new users that not know amiga. The biggest potential are current amiga interested people and people knowing amiga from the past. And 68k is what people know, the biggest software base is there, the most compiler and development environments and so on.

"Not necessarily. There's a reason why a pillar of AROS is source-level compatibility (binary-compatibility is only about 68K)."

I do not understand how this helps... you still need people who adapt software to 64bit and recompile them, SMP-support alone makes no software faster, you need software that uses it. And you need OS-support that makes it possible for software to use it.

Certainly 68k as architecture in the real world is dead but we talk here about a micro market. It would be stupid to compare to one of the big platforms. The platforms we can compare with are exotic platforms (in most cases retro related too). In this niche world 68k as amiga platform is not "dead" (at least it depends how you define dead here). A market is living when it is growing in both number of users and products. This is the case for the amiga 68k market currently.
« Last Edit: December 28, 2020, 12:56:56 AM by OlafS3 »



nikos

  • Senior Member
  • ****
    • Posts: 298
    • Karma: +8/-2
    • aspireos
Reply #47 on: December 28, 2020, 12:34:45 AM
This is how I feel about classic and NG Amiga systems.

Classic Amiga is mainly a gaming, demo scene platform. That is where the system is great. For applications it is mostly not very useful. Maybe for tracking with ProTracker and a few other programs.
For NG systems MorphOS is the most advanced with the best software. What you need there is a dedicated computer for it.
AROS can be run on most computers people have from before. At least hosted or with a VMPlayer.
If you look at how many times AROS been downloaded you will notice there are interest. Even my little distribution is downloaded over 10 000 times.
I'm sure people feel at home after installing AROS. It does for sure feel Amiga, and feels responsive and fast.
I agree with Olaf, and said it myself. It should be a very good Amiga emulator included. That is the first thing people are looking for.


OlafS3

  • Junior Member
  • **
    • Posts: 63
    • Karma: +3/-0
Reply #48 on: December 28, 2020, 12:51:07 AM
This is how I feel about classic and NG Amiga systems.

Classic Amiga is mainly a gaming, demo scene platform. That is where the system is great. For applications it is mostly not very useful. Maybe for tracking with ProTracker and a few other programs.
For NG systems MorphOS is the most advanced with the best software. What you need there is a dedicated computer for it.
AROS can be run on most computers people have from before. At least hosted or with a VMPlayer.
If you look at how many times AROS been downloaded you will notice there are interest. Even my little distribution is downloaded over 10 000 times.
I'm sure people feel at home after installing AROS. It does for sure feel Amiga, and feels responsive and fast.
I agree with Olaf, and said it myself. It should be a very good Amiga emulator included. That is the first thing people are looking for.

I have not really in deep tested Amibridge (only now and then installed Icaros Desktop in virtual environment) so I cannot judge what is already possible and how difficult to install. It would be a good solution if there is not much difference when installing and starting 68k software compared to X86 and the emulation should start and close automatically, at best without being visible at all. On Windows you can define for every application (on Windows 10) that it should run in Win 7 or even Dos. Perhaps something similar could be created, that you define it for every application or game that it runs on A500, A1200 and so on. Just a idea. The screen the programs run in should look like amiga, resources like clipboard should be exhangeable between host environment and 68k software in emulation. Printing of course the same, mounted adf and similar be accessible from host. On the other hand f.e. mounted resources in host environment should be visible in emulated environment (perhaps with different naming). Just a idea how it could be like.
« Last Edit: December 28, 2020, 12:58:00 AM by OlafS3 »



nikos

  • Senior Member
  • ****
    • Posts: 298
    • Karma: +8/-2
    • aspireos
Reply #49 on: December 28, 2020, 01:00:43 AM
This is how I feel about classic and NG Amiga systems.

Classic Amiga is mainly a gaming, demo scene platform. That is where the system is great. For applications it is mostly not very useful. Maybe for tracking with ProTracker and a few other programs.
For NG systems MorphOS is the most advanced with the best software. What you need there is a dedicated computer for it.
AROS can be run on most computers people have from before. At least hosted or with a VMPlayer.
If you look at how many times AROS been downloaded you will notice there are interest. Even my little distribution is downloaded over 10 000 times.
I'm sure people feel at home after installing AROS. It does for sure feel Amiga, and feels responsive and fast.
I agree with Olaf, and said it myself. It should be a very good Amiga emulator included. That is the first thing people are looking for.

I have not really in deep tested Amibridge (only now and then installed Icaros Desktop in virtual environment) so I cannot judge what is already possible and how difficult to install. It would be a good solution if there is not much difference when installing and starting 68k software compared to X86 and the emulation should start and close automatically, at best without being visible at all. On Windows you can define for every application (on Windows 10) that it should run in Win 7 or even Dos. Perhaps something similar could be created, that you define it for every application or game that it runs on A500, A1200 and so on. Just a idea. The screen the programs run in should like like amiga, resources like clipboard should be exhangeable between host environment and 68k software in emulation. Printing of course the same, mounted adf and similar be accessible from host. On the other hand f.e. mounted resources in host environment should be visible in emulated environment (perhaps with different naming). Just a idea how it could be like.

As you say Olaf Amiga is not Amiga. There are different models and configs. There are ECS, AGA and RTG. 68000, 020, 060 etc. etc. On 68k we have whd-load that solves a lot of problems. This program also works under emulation. We even have a whd-load launcher specially made for AROS. To have a solution that is integrated just as running 68k native is not that easy. 98% want to play the old Amiga games and run some demos. AROS is already working  great for that but the 68k emulator should be updated.


OlafS3

  • Junior Member
  • **
    • Posts: 63
    • Karma: +3/-0
Reply #50 on: December 28, 2020, 01:07:18 AM
A whdload-luncher?

Something like that... it should be possible to start a game by simple double clicking without even care that it is whdload and you are on X86. In best case without even needing to launch a program for it. The barriers between X86 and 68k should become as invisible as possible. Also resource sharing is important. There should be no basic difference if you use a 68k application or a X86 application. I know that this is propably not easy to do. I just wrote what I assume would be accepted by more people.



nikos

  • Senior Member
  • ****
    • Posts: 298
    • Karma: +8/-2
    • aspireos
Reply #51 on: December 28, 2020, 01:19:11 AM
A whdload-luncher?

Something like that... it should be possible to start a game by simple double clicking without even care that it is whdload and you are on X86. In best case without even needing to launch a program for it. The barriers between X86 and 68k should become as invisible as possible. Also resource sharing is important. There should be no basic difference if you use a 68k application or a X86 application. I know that this is propably not easy to do. I just wrote what I assume would be accepted by more people.

We all know this would be perfect, but not easy.
Most Amiga games are NODOS so you can not simply click on any game to run it.
This whd-load solved on top of a lot of other things.


cdimauro

  • Junior Member
  • **
    • Posts: 50
    • Karma: +1/-1
Reply #52 on: December 28, 2020, 07:52:26 AM
@cdimauro

That is exactly what I meant with "no clue". That was the case end of 2019 and from this time the post of Gunnar. Kal left all related to vampire if I remember right December 2019. Yes he got a board. The same is true for MSchulz. He got a board and indeed contributed some time but then also left. He has his own project now (a kind of Amithlon on ARM based on Aros 68k). Then Aros was forked, the dev contributing now are involved because they are interested in Aros 68k on Vampire, nothing else.
OK. Thanks for the clarification. I didn't know those details, which weren't publicly available (at least to me, looking at the usual sites which I frequent).
Quote
Kal on the other hand has different views so, in my view, forking was best solution. The forked branch does not affect the main branch, even if something is removed there.
As I've said, if you remove the HIDD to get something similar to the original Amiga o.s. then it's a bad thing, because you're wasting resources.
Quote
And regarding future, of course anything 68k related cannot become mainstream again but in amiga terms the 68k market is multiple of all NG related platforms together regarding sales and number of users. How many active Aros (X86) users are there currently should be known to you.
It looks like that the AROS "market share" is bigger than the 68K one, according to the numbers (and only his numbers) that nikos reported...
Quote
BTW I personal do not see the real potential for another exotic OS out there. It is difficult today to really attract new users that not know amiga. The biggest potential are current amiga interested people and people knowing amiga from the past. And 68k is what people know, the biggest software base is there, the most compiler and development environments and so on.
Which can run under an emulator, if needed.
Quote
"Not necessarily. There's a reason why a pillar of AROS is source-level compatibility (binary-compatibility is only about 68K)."

I do not understand how this helps... you still need people who adapt software to 64bit and recompile them,
If you follow the guidelines you don't need explicit support for 64-bit software (as well as for the endianess): it should be enough to recompile.

However some bug can be left in the code, and it can eventually be easily fixed.
Quote
SMP-support alone makes no software faster, you need software that uses it. And you need OS-support that makes it possible for software to use it.
On AROS there's already o.s. support for SMP, and even allowing single applications to run concurrently (e.g.: not requiring explicit SMP support. Just launch them, and let the system assign them to one of the available core / hardware threads) it's already a vast improvement compared to the single-process-for-all-threads Amiga o.s. model, especially nowadays where even cheap phones have several cores.
Quote
Certainly 68k as architecture in the real world is dead but we talk here about a micro market. It would be stupid to compare to one of the big platforms. The platforms we can compare with are exotic platforms (in most cases retro related too). In this niche world 68k as amiga platform is not "dead" (at least it depends how you define dead here). A market is living when it is growing in both number of users and products. This is the case for the amiga 68k market currently.
The Amiga 68K "market" is made by the same people which are splitting their interests on one or more of the available "post-Commodore" platforms: AROS, MorphOS, OS4, FS/WinUAE, Minimig/Mist/Vampire/FPGA*.

The only ones from those which are used as a normal platform aren't the 68K ones, because they have not enough processing power, and their are mostly used to run videogames, where compatibility is the biggest needed feature, and WinUAE wins hands down.

Talking about applications, which is the second use case, the non-68K platforms are the most widespread (because you need more resources / performances).

So, the Amiga 68K market is mostly a retrogaming platform: insert the "disk" and play. And it has no future other than that.

68K is dead because it's a legacy platform, and the fact that there's a "FrankenstIEn" (cit.) being actively developed doesn't mean that it's back alive, or will have any chance in the future to increase the market share of the whole Amiga nano-niche.
Even because it's ultra-expensive for what you get, whereas a MiniPC costing a few bucks can give better results in terms of both compatibility (you can configure WinUAE as you want, contrary to all hardware solutions) and performance. There's no point on getting an FPGA/hardware-devices other than people which psychologically needs to "touch the physical hardware" (SIGH!).

A whdload-luncher?

Something like that... it should be possible to start a game by simple double clicking without even care that it is whdload and you are on X86. In best case without even needing to launch a program for it. The barriers between X86 and 68k should become as invisible as possible. Also resource sharing is important. There should be no basic difference if you use a 68k application or a X86 application. I know that this is propably not easy to do. I just wrote what I assume would be accepted by more people.
There always will be differences because you cannot fully integrated 68K and native apps. This can only happen on big-endian & 32-bit systems, like OS4 and MorphOS are doing.

However sharing some resources (filesystem, clipboard, networking, etc.) and intercepting and launching 68K apps or WHDLoad games in a transparent way should be enough.

The best in my opinion is realizing a "virtualizer", like VAMOS, and allowing to run 68K apps transparently on any host o.s., but this requires time as well.



magorium

  • Senior Member
  • ****
    • Posts: 253
    • Karma: +7/-0
  • Convicted non contributor
Reply #53 on: December 28, 2020, 08:45:12 AM
Are we going to continue from the old aros-exec forums in that every thread is hijacked with nonsense about ABI's/API"s and whats more, and what or shouldn't be done by the developers etc ?

It was tiring then and even more tiring now and for sure clouds the question related to the original poster. Can't you just start your own wank- or distortion- channel or something ?


cdimauro

  • Junior Member
  • **
    • Posts: 50
    • Karma: +1/-1
Reply #54 on: December 28, 2020, 12:25:02 PM
Please, can a moderator move the above posts in the following thread: https://ae.amigalife.org/index.php?topic=647.msg5707#new ?



paolone

  • Senior Member
  • ****
    • Posts: 448
    • Karma: +13/-0
Reply #55 on: December 28, 2020, 12:40:26 PM
I have not really in deep tested Amibridge (only now and then installed Icaros Desktop in virtual environment) so I cannot judge what is already possible and how difficult to install. It would be a good solution if there is not much difference when installing and starting 68k software compared to X86 and the emulation should start and close automatically, at best without being visible at all. On Windows you can define for every application (on Windows 10) that it should run in Win 7 or even Dos. Perhaps something similar could be created, that you define it for every application or game that it runs on A500, A1200 and so on. Just a idea. The screen the programs run in should look like amiga, resources like clipboard should be exhangeable between host environment and 68k software in emulation. Printing of course the same, mounted adf and similar be accessible from host. On the other hand f.e. mounted resources in host environment should be visible in emulated environment (perhaps with different naming). Just a idea how it could be like.


I like when people insist describing what Amibridge is all about, in terms of "there should be something doing this, that, and that more".


paolone

  • Senior Member
  • ****
    • Posts: 448
    • Karma: +13/-0
Reply #56 on: December 28, 2020, 12:44:52 PM
@cdimauro


Apollo team's 68K version of AROS will be a special fork of AROS aimed at replacing AmigaOS at every level, it won't require portability since it will work on some 68K architectures only, with original custom chips and Vampire expansions. So there is no need for the hidd system and for the HAL, which will still be included with main AROS system and ports to other architectures, as usual.