AmigaOS 3.1.4 withdrawn from market after Amiga

[Bruce Abbott]

Originally Posted by gregthecanuck

I think that open-sourcing at least some parts of OS3.1 could really help.

So people can mess with it and introduce more bugs and incompatibilities? No thanks. I would rather have an official closed source controlled by developers who know what they are doing.

Pick something. scsi.device?

Anybody can write their own device driver - no need to have source code for the 3.1 ROM version. Or are Amiga developers too lazy to write their own code now?

I was going to get 3.1.4, but decided to wait and see how it developed first. I'm glad I did because it doesn't offer anything I really want and some of its 'features' are dealbreakers for me. I'll just stick with 3.0 since it does everything I need.

[gregthecanuck]

Originally Posted by Bruce Abbott

So people can mess with it and introduce more bugs and incompatibilities? No thanks. I would rather have an official closed source controlled by developers who know what they are doing.

Anybody can write their own device driver - no need to have source code for the 3.1 ROM version. Or are Amiga developers too lazy to write their own code now?

I was going to get 3.1.4, but decided to wait and see how it developed first. I'm glad I did because it doesn't offer anything I really want and some of its 'features' are dealbreakers for me. I'll just stick with 3.0 since it does everything I need.

Hi Bruce -

My suggestions on open source suggest a similar model that is used elsewhere - there is a main repository that nobody can update except the "gatekeeper(s)" (such as Mr. Torvalds on Linux). That way developers can fork their own branch, hack away on suggested enhancements/tweaks, and feed them back to the main branch for inclusion by request (a "pull request"). This type of model is complemented by a healthy forum or mailing list discussion.

I suggested scsi.device since from previous notes from the developers it is full of a lot of old classic crud. Maybe it could be rewritten in public? That would unencumber the code from the "no sharing the code" restriction in Hyperion's agreement with A-Inc.

As for your other comments, well, everyone has their opinion.

[kolla]

Originally Posted by Bruce Abbott

So people can mess with it and introduce more bugs and incompatibilities?

Yes, and force you upgrade at gun point, or something.

I'll just stick with 3.0 since it does everything I need.

The version with most bugs and compatibility issues - good choice!

The point of open sourcing would be for academic reasons mostly, and for freedom for anyone to read code and learn without the risk of "legal prosecution" or being "locked out" from developing for example on AROS.

Have you noticed all the new hardware for Amiga lately? New boards etc? Ever thought it might have had something to do with the leaked 3.1 sources from a couple of years ago? People learnt from them.

[gatti]

amigastore.eu is working on it, and hope it will be its next news (source: twit from official account)

[Locutus]

Originally Posted by kolla

Ever thought it might have had something to do with the leaked 3.1 sources from a couple of years ago? People learnt from them.

I thought we had all been told they would be utterly impenetrable.....

[daxb]

As far as I know the 3.1 sources where available (leaked) ~20 years ago. So a long time learning.

[kolla]

Yes they were, that's how they were "re-leaked" as it were. Availability improved a lot a couple of years ago though.

[wawa]

Originally Posted by Bruce Abbott

So people can mess with it and introduce more bugs and incompatibilities?

this is possible with a closed source as well. whether the project is successful is a matter of responsibility and consequent involvement. it matters for instance that there is enough (quality) feedback between developers and users/testers (both ways) which implies that there are enough and as able as motivated people involved.

[wawa]

Originally Posted by daxb

As far as I know the 3.1 sources where available (leaked) ~20 years ago. So a long time learning.

which (i think we are agreeing here) indicates, that it might not change much, if the sources were legally open. this is simply too much work for any single person. you need an infrastructure for that kind of huge project to maintain and work on it. such an infrastructure exists on aros. so far i see thor and olsen must have at least started to introduce this sort of infrastructure to their os/kickstart sources, as they must have been totally outdated in this respect, as they reported. it also needs a perspectivic wholesale approach, instead hacking some bits and pieces of asm together, and im sure they have proved to think in these categories. thats what others might not have considered and why leaking the sources couldn't have any such effect.

[kolla]

The last "leak" was at new-year of 2016 and a few months later a certain piece of software came out that has been become the de-facto tool to help people build Amiga hardware. Coincidence? You decide!

[Romanujan]

What piece of software?

[Locutus]

What i found most interesting was that back when the 3.1 code dropped it was the first time in ages that i've heard so many people excitedly talk about the Amiga.

Real people in person having interesting in depth discussions about how stuff works on the Amiga in loads of places.

Far more interesting then anything else in the last 10 years.

[Mr.Flibble]

I guess we're not going to see much happen with all of this officially until the legal process moves onward.

[Pyromania]

I want to be able to buy 3.1.4 in the next two weeks.

[Bruce Abbott]

Originally Posted by kolla

Yes, and force you upgrade at gun point, or something.

Yes, that is effectively what happens.

When the OS was distributed on mask-ROMS inside new computers there was a strong incentive to keep bugs down and the release frequency low. A few machines such as the A1000 and early A3000's had the ability to 'soft-kick', so Commodore distributed beta kickstarts to developers for testing. But both developers and users only had to deal with a small number of release versions which were carefully crafted for backwards compatibility and had well defined differences.

Contrast that with disk-based OS's on other platforms. Linux has so many different versions it's a joke. How are users supposed to deal with it? 'Just' compile your applications from source, or even recompile the kernal! (and pray it still works). If AmigaOS is open-sourced then we may soon be the same boat.

The version with most bugs and compatibility issues - good choice!

Also the version with the least bugs and compatibility issues! When there is only one...

The point of open sourcing would be for academic reasons mostly, and for freedom for anyone to read code and learn without the risk of "legal prosecution" or being "locked out" from developing for example on AROS.

AmigaOS API's are well defined so what's the point of examining the source code?

Have you noticed all the new hardware for Amiga lately? New boards etc? Ever thought it might have had something to do with the leaked 3.1 sources from a couple of years ago? People learnt from them.

A lot more 3rd party hardware was produced before 3.1 was leaked. How did they manage it without learning from the source code?

[kolla]

Originally Posted by Bruce Abbott

Contrast that with disk-based OS's on other platforms. Linux has so many different versions it's a joke. How are users supposed to deal with it? 'Just' compile your applications from source, or even recompile the kernal! (and pray it still works). If AmigaOS is open-sourced then we may soon be the same boat.

It's already in the same boat, consider all the patches and unofficial updates people need to use to make even plain OS3.1 work on just about any Amiga hardware.

AmigaOS API's are well defined so what's the point of examining the source code?

To find out how it treats hardware, for example.

A lot more 3rd party hardware was produced before 3.1 was leaked. How did they manage it without learning from the source code?

When CBM was alive, there was CATS and there were developers to consults with, they provided all the help hardware manufacturers needed, including source code if necessary.

[Bruce Abbott]

Originally Posted by kolla

consider all the patches and unofficial updates people need to use to make even plain OS3.1 work on just about any Amiga hardware.

Considering... it's been a long time since I had KS3.1 on an A3000, A600 and CD32, but I don't remember having to do 'all the patches and unofficial updates' you say people need to make even plain OS3.1 work. I just used the stock OS 3.1 system disks and everything worked fine. I even put the KS2.1 ROM back in my A600 and it mostly worked fine though it still had WB3.1 on the hard drive.

To find out how it treats hardware, for example.

The one time I 'needed' to do that I just disassembled the relevant part of the ROM. But that only told me what the hardware was expected to do, ie. no more than I already knew from the Autodocs etc. I bet the source code wouldn't have have provided any more insight.

When CBM was alive, there was CATS and there were developers to consults with, they provided all the help hardware manufacturers needed, including source code if necessary.

I remember CATS, but I was only a software developer so perhaps I wasn't in on it. However I don't remember Commodore offering the full OS source to anyone who asked - and I'm pretty sure they wouldn't hand over any of it without an ironclad NDA. Which is the way it should be. Developers should be able to get all the information they need from officially published documentation without having to trawl through the source code!

The real reason people want AmigaOS open-sourced is obvious - they want to hack it to do what they want rather than go to the effort of developing code which works with the stock OS. And we all know what the result will be - dozens of customized OS versions each with its own peculiarities, and lots of frustrated users having to constantly change their OS just to run other people's stuff. It's bad enough now with what seems like 90% of new Aminet uploads not working on stock machines...

[alpine9000]

Originally Posted by Bruce Abbott

It's bad enough now with what seems like 90% of new Aminet uploads not working on stock machines...

What’s “a stock machine”?

I made my most recent game work on I think (didn’t test below 1.3) all Kickstarts, what a pain and it barely uses any OS functions.

For me the reason it would be valuable for the OS to be open sourced is to ensure the exact topic of this thread doesn’t drag on forever until finally the rights end up frozen in an estate that doesn’t even know what it owns.

[modrobert]

Originally Posted by Bruce Abbott

The real reason people want AmigaOS open-sourced is obvious - they want to hack it to do what they want rather than go to the effort of developing code which works with the stock OS.

For me the reason to go open source would be to collaborate as a community, fix bugs, add functionality in the form of maintained forks, and perhaps most important; having a common open code base to reference during discussion (quote examples, reference system calls, etc.).

Several of the members here at EAB are very skilled at reversing any 68k binary code you throw at them, obfuscated or not, hacking AmigaOS does not require source code.

[kolla]

Originally Posted by Bruce Abbott

Considering... it's been a long time since I had KS3.1 on an A3000, A600 and CD32, but I don't remember having to do 'all the patches and unofficial updates' you say people need to make even plain OS3.1 work.

So you didn't have a 060, graphics card, usb host controller or large hard drives, did you.

I didn't say that CBM gave away entire source code, but relevant parts, under NDA. One of the big controversies with MorphOS was regarding this, as CBM Germany had handed out OS 3.1 sources to Phase5 at some point.