Some honest words

[RC_tech]

Hello!

I'm starting this thread to check out the status of AROS and what is actually going on. Some people may find some things offensive. I'm adding numbers to the topic, so use them if you're replying.

To me: I started with 1989 with a C64 and got my A500 in about 1991. I bought my A4000T new for 3999 DM and got very soon the CV64/3D and my Cyberstorm MKII/060. I'm a complete classic user and don't like FPGA stuff. That's why I have for my C64 two CMD HD's, one FD-4000 and the holy grail - a SuperCPU 128. I got it for just 450 Euro ;-) No SD card or disk image was ever mounted on the 8 bit machines. I'm also since March 2006 in the OSx86 scene.


The Amiga scene is split in a bunch of incompatible versions: Original 68k AmigaOS, AmigaOS 4, MorphOS, AROS x86/64, AROS PPC, AROS ARM.

1. Check out rationally: Is AROS PPC and AROS ARM really neccessary? The PowerPC processors are a dying platform. At least MorphOS can use old PPC-Macs which are looking nice, but if they are defective there are no spare parts, and it's difficult to find somebody who repairs them. Real AmigaOS 4 machines are much too expensive. And ARM? I tested the Android hosted version. Nice, but what about software and a working onscreen keyboard? For ARM we have Uae4All, this one is working very fine. It would be best if all work is concentrated to the x86/64 version.

2. No operating system works without drivers. And there's not even an attempt to support modern hardware. I'm not talking about multi-core processing or 32 Gigs of RAM. Mainly graphics is the problem. I's a very bad sign that there is only 3D support for a few old Geforce cards. No AMD or Intel. I'm sure there are a lot of people in here who want to get a cheap laptop for AROS. If AROS would support the integrated HD Graphics, this would be a large step forward.

I know there are only a few developers. But AROS is designed to compile 68K AmigaOS / Hardware to work on x86 PC Hardware. Wouldn't it be a task for operation system compiling / combining? The idea is to compile linux software to AROS software. This would make the creation of AROS drivers much simpler. OSx86 also got a lot of driver conversions from Linux. Maybe it's possible to create a library which can make use of linux drivers more easily. I'm thinking on something like the rexxbridge.library: It was an addon for ARexx to access all other library, even intuition.library.

3. Harddisk support. The AmigaOS RDB is the best partition block I know. But it doesn't make sense on a x86 platform, you have to use MBR in any case. My recommendation would be to keep AROS/RDB for old computers, for newer ones GPT should be used. In single partitions, not a RDB in a GPT partition. This also would allow multiple bootable AROS versions. The blocksize is fixed to 4K, because all modern harddisks and SSD's are using 4K sectors, not 512 byte. To make AROS partitions more compatible to other operating systems, a common file system should be used, either NTFS or ExFAT. Additional data which were usually stored in the RDB can be stored in another file in the root directory, f.e. diskdata.info. Somebody who also uses OSx86 should check out if it's possible to integrate AROS into Clover. This would allow a much easier multiboot with other operating systems.

4. Software. There is a lot, but what? Some games, but most machines have no 3D support. Software like Blender isn't using everybody and also needs 3D support. We need LibreOffice, Gimp and similar tools. And a conversion of Firefox. Even a version for old PowerPC Macs is available, 'TenFourFox'. There are seperate versions for G3, G4, G4+ and G5. Why is there no Amiga/AROS/MorphOS version available? Simple: Amiga developers always want money. 'Bounties' are very common in the Amiga scene. But why donate if you have to fear if the project will never be finished? In the OSx86 scene eveything is free. In Insanelymac the word 'TonyMacx86' is banned just because his software isn't open-source. I'm heavily against paying since the very early OSx86 developer JaS set up a web page 'Donate for my Macbook Pro'. As he got it, he left the scene. And the last install DVD he did, 10.4.8, had a heavy fault in the Disk Utility. That's why I'll never pay to a single person.

5. Icaros Desktop. You're doing a great job with this, Paolo. But there's no noteworthy alternative. If Icaros doesn't work, nothing will work. Often enough the Live DVD crashes during boot, no matter what option was selected. This can be already a problem because some drivers will fail because the hardware isn't suitable for the DVD's configuration. In early OSx86 times were attempts to create OS X Live DVD's using the NetBoot feature, but of course it only worked on a very few hardware combinations. This problem could be solved if there's an installation-only mode for Icaros. Android-x86 is programmed by one person and the downloadable ISO uses a live system for testing and a text-based installer, it looks like MS-DOS. Being as close to the PC hardware could at least allow an installation. The installers also should be updated. For example, the installation of graphics drivers. Actually, only nVidia works. It's possible to check PCI ID's. If a graphics card with the vendor ID 0x10DE (nVidia) exists, noveau.hdd will be installed. The same is for WLAN, network and audio. Sure, this is more complicated, but it makes the finished installation more compatible and is more user-friendly. The optional installation of the 64bit-Kernal would also be great. Is it possible to add the Icaros ISO on an USB stick, f.e. with WinSetupFromUSB? It works for Linux, Android, FreeDOS and for many other operating systems, but the last time I checked, it didn't work with Icaros.


So, this is for now. I would like to have a good working AROS system, because I like the Amiga. Programming in AmigaDOS and Arexx is real fun. But setting up an old computer with a crappy nVidia graphics card just for AROS wouldn't be a good idea. Every operating system has to go with modern hardware. Windows and macOS always got worse. OS X 10.6 is awesome fast, but it had not much unneccessary iOS stuff, it was an operating system for a computer. Since 10.11 Apple even blocked access to system folders with the 'rootless mode'. Android 4.4 blocked access to the external SD-card. Very good idea, to have a device with 8 Gigs internally and a 128 GB SD-card which can only be read. I don't know about Linux, I don't think it got much worse. So, it would be a good idea to make AROS better than as it is now. The most important things to change are listed again below.

- Concentrate developing on AROS x86/64.
- 3D drivers for Intel HD graphics at least since HD2000, second generation Core i.
- Maybe a method to compile Linux drivers to AROS.
- No RDB on modern computers, using a single GPT partititon for each AROS partition.
- Switch to NTFS or ExFAT as main filesystem.
- Icaros Desktop PC textmode installer for more compatibility.
- Installer with hardware detection, but manual selection.

Greets, RC.

[trekiej]

I believe what you are saying is quite known to this community.
I do like what you are saying. I wish to go to Aros 64 bit too.
Getting more developers is the problem. It would be nice to support one platform and see it move ahead.
The ARM port is suppose to bring low cost hardware, in my opinion.

[cavemann]

I run AROS on x86 and use Icaros, so I can feel your pain. But you must understand that many people are more interested in AROS 68k and Arm, than in the x86 branch.

Modifying Linux drivers has been suggested before. But as you said, AROS does not have the devs to do that, even if it was the right course to take.

NTFS is an interesting thought. I would prefer the Amiga style Pfs3. But the devs are focused on other priorities. And most likely, we won't see either soon.

Intel graphics updates. I've been told that 3d for those drivers cannot be updated and would have to be meticulously rebuilt from scratch. And the software you mentioned would need some libs overhauled.

As you know the devs have not stabilized 64-bit. Even without the changes you seek, it would take some time to transition the software to be on a par with Icaros. And Icaros keeps getting better 

But I must agree, at one time the idea of AROS was to run native on common hardware, and it seems like we may be losing some of that.

[OlafS3]

@RC_tech

Regarding ARM I disagree with you

In opposite to you I think ARM (here expecially RPi port) has the potential to bring new users

BTW the current two most active aros branches are ARM (RPi) and 68k (for real hardware like vampire)

[RC_tech]

And this is why I don't understand: The Raspberry Pi is a not very powerful machine. AROS has less native software than AmigaOS and it's not useable for gaming. UAE4all is a good working emulator and it's available as a standalone operating system for the Pi. And 68k? I don't like the Vampyre because it's FPGA stuff and it's way too expensive. Does it come with AROS preinstalled? How compatible is AROS to the original Kickstart? I'd see no reason to set a AROS 68k ROM into my classic Amiga hardware.

@cavemann:

The same: I really don't understand why ARM and 68k is so important.

There are only a few developers. But Android-x86 is programmed by just one person, and the actual version is 8.1. So, it is possible to create and maintain an operating system with only a few developers. GEOS 64/128 also had just a bunch of official programmers at Berkeley Softworks, but here were lots of C64/128 existing and the OS Version 1.2 and 1.3 were bundled with computers, I think 1.5 was in the package with the 1351 mouse. This is a rather similar idea like today: Give out an operating system for free for a widely available platform, with detailled programming information to get more software. And it worked, in it's high time GEOS was the most popular operating system worldwide - just behind MS-DOS and MacOS Classic.

There are enough reasons why AROS doesn't get developers: It's based an old-school operating system with lack of software, no drivers and a split community which isn't preparing for the future.

Programming drivers from scratch is not a reason not for trying. In the OSx86 scene this was often done, because Apple changed some parts of the OS during newer versions, so old drivers weren't working anymore. The bootloaders Chameleon and Clover were also written from scratch, they're not a modified version of GRUB or so.

x86 is the biggest platform for desktop computers and laptops. If AROS don't want to support this platform, it has absolutely no future. But there could be a method. As Apple released OS X Intel, it required the SSE3 command set. Some CPUs like the Pentium 4 or early Athlon 64 processors didn't have it. So the first patched OS X kernal (10.4.4) had an SSE3 emulator which added the new commands. AMD and Intel processors always had differences, AMD needed manual patches. Today this is done on the fly by routines of the bootloader. So, if ARM is one of the major branches, why not include an ARM emulator? With this method, it would be possible to use ARM software for AROS. But if there's said from the beginning: Nooo, there are to few developers and we have to start from scratch, then this is a sign that there is no interest in AROS-x86.

An ARM emulator could also be a base for an Android App compatibility. They're written in Java versions, it should be possible to implement it. Even if the Play Store is not available, installing an .apk should be possible. But oh well, a lot of Android Apps need 3D support, so we're back to the main problem.

[OlafS3]

@RC_Tech

And who will develop the drivers for all the different hardware?

RPi is of course not the fastest hardware but Michal wrote that he would also support faster ARM hardware after RPi

RPi (for now) has several advantages:
it is relativ static hardware so it can fully supported
there is a big community open for exotic options behind it
and the obvious one... it is very cheap

FPGA hardware might be not for you but it is obviously interesting for many others. And if you are not interested in 68k branch of Aros then it is your decision.

[OlafS3]

BTW of course Aros is open source so basically everyone with some skills can contribute

So when you miss drivers or want development on the 64bit X64 branch then do it

If you can not then you must accept what developers decide for themselves

[Marlon]

Work is being done to get ABI-v1 completed this year. There is work being done on x86/x86_64 as well as ARM-BE and 68k. It's just that ARM-BE and 68k is much closer to "completion". Going forward and just push out x86_64 before ABI-v1 is even complete, just for the sake of getting it out there is just plain and simply dumb. It will come. The work being done for 68k and ARM-BE reflects for the whole ecosystem of AROS ABI-v1 as a whole.

x86_64 will get some love too, just be patient. There's like 2 main developers doing it all on their spare time. And they work on what motivates them the most, can you blame them? I certainly don't. Personally I welcome AROS ARM-BE with open arms (no pun intended). The goal is to support to RasPi 2 & 3, Pine64/RockPro64 and the eventually the PineBook https://www.pine64.org/?page_id=3707

That would satisfy the need for an AROS laptop and would circumvent the need to support the countless of PC-chipsets and different configurations in the PC laptops.

With the amount of devs AROS has today PC is not really a viable target because of all the different configurations there is. ARM is an easy target because it's pretty static and there isn't thousands of configurations (yet). 68k is also an easy target as I don't see the Amiga Classic computers changing much anytime soon. Getting these targets completed first means that there will actually be at least 2 targets that are actually complete and usable.

You can say "Do this and that, because of this", but it won't change anything. The situation is what it is. You either pick up the tools and help out, or just deal with it.

[trekiej]

Come on 64 bit!
Not that 32 bit is bad or lacking necessarily.

[wawa]

what do you want? there is x86_64 target for aros and it works afair, there is even x86_64 smp target. ppc is not being maitained very actively right now, since there aint much interest from existing devlopers and apparently not a lot interest from ppc developers. but it is being dragged along, and likely at least partly profits from the fixes done for other platforms. so what are you complaining about?

[paolone]

Hello!

I'm starting this thread to check out the status of AROS and what is actually going on. Some people may find some things offensive. I'm adding numbers to the topic, so use them if you're replying. And blah blah blah.

Once upon a long ago there was a nice but heavily undeveloped operating system I loved so much, which was missing a proper distribution. I wonder what would have happened, if I just sent a message to AROS-EXEC to complain, instead of making one.
And I am no developer at all.