Official update to OS 3.1 from Hyperion

[ascp]

Originally Posted by Korodny

Okay. You downplaying any statements regarding the complexity of such an effort (working on the existing code) in this and the other thread made me wonder.


Would you mind telling us who that would be? I've been following this very closely for a decade and a half (I'm covering it for amiga-news.de), and I certainly have no "pretty clear idea" who might own these materials. Could be either ESCOM's creditors or Amiga Washington's, but that are just wild guesses.


There is no research yet. I love the Amiga Documents effort, but it's not a substitute for proper research. In the trial against Hyperion even the judge noticed at one point that the AInc crowd was simply making up documents as they needed them (they had backdated an allegedly for or five year old contract a week too far) And the players from that trial are dead or gone now, Same goes for the original AInc investor ("Invisible Hand", IIRC) that was shut down ages ago.

The whole point of that entire AInc Washington/KMOS/AInc Delaware mess was to make sure there's no trail to follow. These were all privately held companies, so the only ways to extort information would be to find creditors willing to be part of a trial and fund them.


Again: unless you're willing to burn a lot of money, the only sane approach would be to not look too closely, assume that nobody else will because of the problems listed above and make an offer to Cloanto.


The Amiga is a retro toy now, 99% of the user base has no real need for continued development - they certainly wouldn't mind, many would probably even donate to projects or buy ready-to-use ROM chips. But if none of that happens, they'll be fine with the ready-to-use ROM chips and Kickstart images available now and the P96 archive from Aminet.

You're simply 20 years to late.

One brutal effort would of course be to declare the sources public domain, and then let any right holders prove in court that the source are theirs. That way, either the real right holders would have to emerge with proof or the source would be in public domain - that would be up to the court to decide.

[ovale]

Hello, I would go for a super liberal license like MIT or BSD.

GPL will make impossible to statically link code which is not GPL compatible (see Linux and ZFS as an example). Static linking is even more important than in Linux in a limited platform like 68k Amiga.

A license like MIT is a no brainier. GPL always requires considerations.

For example AROS license is described as not GPL compatible. This will make hard to mix code between projects.

[Korodny]

Originally Posted by ovale

GPL will make impossible to statically link code which is not GPL compatible

Statically link what to what? Nobody's statically linking OS modules to his own code. But even if that was a thing: as long as you stay binary compatible - which would obviously be the case - people could always link to some older (proprietary) version of a library.

I would go for a super liberal license like MIT or BSD.

So, you'd free the OS using a massive amount of resources - and then enable people to close it down again? That doesn't make much sense, IMHO.

GPL always requires considerations.

Which is actually a good thing, because it keeps the likes of Cosmos out.

For example AROS license is described as not GPL compatible. This will make hard to mix code between projects.

IIRC, The AROS license is simply the Mozilla Public License with the term "Mozilla" globally replaced with "AROS" - an extremely stupid move IMHO that should be reconsidered as soon as possible.

The Mozilla Public License has a clause always allowing Code to be relicensed under later versions of the same license, and MPL 2.0 finally introduced GPL compatibility. Something similar should be possible with the AROS license, along with finally fixing the license's name for good.

[wawa]

Originally Posted by Korodny

So, you'd free the OS using a massive amount of resources - and then enable people to close it down again? That doesn't make much sense, IMHO.

christoph?

actually my understanding is, it allows people to fork that open source and keep this fork closed, the opened source would remain open, and the changes done to forked source files might be required to be kept open, as it is in case of aros license.

however stupid modzilla or aros license might be, it has not led to close any of their sources by now. instead, it allowed closed source spinoff attempts alike arix, which even if effectively led to nothing themselves, have not had any negative impact of aros itself. rather to the contrary, they were first to have proven that multiprocessing is possible with amigalike os, had some code ported back and today aros itself can follow in their footsteps, even if the implementation is different.

[Korodny]

Originally Posted by wawa

christoph?

Yes?

actually my understanding is, it allows people to fork that open source and keep this fork closed, the opened source would remain open

Exactly. But that's still a problem if some party pays developers to add stuff and keeps the result to itself - you'd still end up with a closed source product that people might want.

Not saying this is likely, but if you've suffered from the negative effects of proprietary software was much as we did, using the GPL (or, more appropriately, the LGPL) makes a lot of sense, IMHO.

however stupid modzilla or aros license might be

The Mozilla license is not stupid. Taking it and renaming it is - because it creates all sorts of license compatibility hurdles.

[wawa]

Originally Posted by Korodny

Yes?


Exactly. But that's still a problem if some party pays developers to add stuff and keeps the result to itself - you'd still end up with a closed source product that people might want.

Not saying this is likely, but if you've suffered from the negative effects of proprietary software was much as we did, using the GPL (or, more appropriately, the LGPL) makes a lot of sense, IMHO.

why is that bad? if people are satisfied with it, then fine. commercial development doesnt have to be necessariy wrong, or some combination of commercial and free initiatives. im completely okay with os4 or morphos or eventually other closed options, like arix would be, as long as im not being forced to support it for the sake of the platform.

therefore i think most welcome license by majority of the parties would be lgpl. but it is moot point, since nothing will happen.

The Mozilla license is not stupid. Taking it and renaming it is - because it creates all sorts of license compatibility hurdles.

dunno why they did it. but i think if it is a problem, then rather a minor one.

[wXR]

For what it's worth, I am not okay with proprietary software. Where my efforts are concerned, I will push for a GPLv2 relicensing. As Korodny himself begins to suggest, I also don't ever want to see any future form of it closed or exploited by anyone again. GPL may seem brutal in some respects, but I don't think it is without justification.

@jonathan

I had proposed your "declare it public domain" idea elsewhere, strongly suggesting to everyone that none of the companies involved in this had enough legal resources -- and possibly none of the rights -- to pursue anything other than cease-and-desist letters. I wanted everyone to channel their inner pirate, or at least their inner courage, and work on the 3.1 sources in public. However, it seems that very few folks who are still left around here are willing to do that. So going down this horrible road is pretty much the only option left, in order to be inclusive.

@Korodny

I'm appreciating now some of the nuance in the point you were making about not looking too hard. And perhaps this is sufficient.

[Acill]

Okay so I just used the byteswap option on my burner software on the A500 version and its fine, but not sure about how to burn the double ROMS for my A1200 as an example. ROMsplit doesnt work on them.

[Olaf Barthel]

Originally Posted by wXR

For what it's worth, I am not okay with proprietary software. Where my efforts are concerned, I will push for a GPLv2 relicensing. As Korodny himself begins to suggest, I also don't ever want to see any future form of it closed or exploited by anyone again. GPL may seem brutal in some respects, but I don't think it is without justification.

@jonathan

I had proposed your "declare it public domain" idea elsewhere, strongly suggesting to everyone that none of the companies involved in this had enough legal resources -- and possibly none of the rights -- to pursue anything other than cease-and-desist letters. I wanted everyone to channel their inner pirate, or at least their inner courage, and work on the 3.1 sources in public. However, it seems that very few folks who are still left around here are willing to do that. So going down this horrible road is pretty much the only option left, in order to be inclusive.

@Korodny

I'm appreciating now some of the nuance in the point you were making about not looking too hard. And perhaps this is sufficient.

So, how will step #2 look like, once funding has been secured and code has been licensed under the GPL?

I've been around for a while, and every time community efforts materialized to forge a path for the Amiga operating system to be liberated, things stalled at the first step (I still have the big "Team Amiga" and "Phoenix" mailing list contents in storage, from almost 20 years ago). As much energy as you can expend on breaking the chains which hold the operating system, the next difficult part is in figuring out what to do about it once it's been liberated.

This is still a fairly large lump of code, not in the kind of shape which allows you to build and maintain it quickly. The tools which the operating system were created with are still as old as the last AmigaOS release version. Worse still, these old tools are still better adapted to the task than the ones you could apply today (e.g. GCC for 68k).

My point is that the first step, acquiring and licensing the code, will fall flat unless you also spend effort on figuring out how to make use of the operating system code. For that you both need an organization to tend to it, and the people with the knowledge to make that happen. Neither of this is going to come together on its own accord.

Making software for the Amiga, maintaining it, is not a growing field. While you can conveniently find copies of the development material and documentation, the breadth and depth required to become familiar and proficient with it is not easily attainable.

Access to the operating system code is important, but so is the other half of the equation, i.e. what to do with the code, and how to enable that. I believe that without covering both ends, the results will be more than just disappointing.

[michaelz]

Originally Posted by Olaf Barthel

So, how will step #2 look like, once funding has been secured and code has been licensed under the GPL?



I've been around for a while, and every time community efforts materialized to forge a path for the Amiga operating system to be liberated, things stalled at the first step (I still have the big "Team Amiga" and "Phoenix" mailing list contents in storage, from almost 20 years ago). As much energy as you can expend on breaking the chains which hold the operating system, the next difficult part is in figuring out what to do about it once it's been liberated.



This is still a fairly large lump of code, not in the kind of shape which allows you to build and maintain it quickly. The tools which the operating system were created with are still as old as the last AmigaOS release version. Worse still, these old tools are still better adapted to the task than the ones you could apply today (e.g. GCC for 68k).



My point is that the first step, acquiring and licensing the code, will fall flat unless you also spend effort on figuring out how to make use of the operating system code. For that you both need an organization to tend to it, and the people with the knowledge to make that happen. Neither of this is going to come together on its own accord.



Making software for the Amiga, maintaining it, is not a growing field. While you can conveniently find copies of the development material and documentation, the breadth and depth required to become familiar and proficient with it is not easily attainable.



Access to the operating system code is important, but so is the other half of the equation, i.e. what to do with the code, and how to enable that. I believe that without covering both ends, the results will be more than just disappointing.

The discussion is going all over the place, I think the right place right now is the thread "Critical components to open source".

[wXR]

@Olaf

Thank you for your thoughts.

I'm offering money and some leg-work to get the process started, I really don't have the knowledge, skills, or frankly the time, to take it any further than that. However, it is in fact a step in a sequence, is it not? The GPL relicensing can hardly "fail" if the GPL relicensing is the goal itself. As you point out, it is just one step of many (perhaps) required to manifest some idea of progress around what is currently known as AmigaOS, but it is at least a concrete goal in that one respect.

I am only offering to help with this first step. My belief is that this puts the entire community in a stronger position, because suddenly the freedom to do something will at least be there. How could that be any worse than the current situation? I can't really understand any arguments to that effect. If an "organization" is required to go further, someone can create one. Personally I am not going to do that, as I am not the all-around Amiga messiah -- and thank god! To be clear, I'm just a guy, and I don't have any such plan. On the other hand I would argue that the supposed rightsholders don't either...

[idrougge]

Getting the entirety of AmigaOS 3.1 to build from source is far from easy going by Olsen's accounts, especially if you intend to make a 512 kB Kickstart out of it.

However, an opened source means that
1) bugs and missing features in separate components, such as a single library or C: command, can be fixed
2) said functionality can be reproduced in a new implementation with less risk of failing on edge cases or wasting time trying to reverse-engineer intended behaviour
3) sooner or later, someone with too much time may be able to make it build when said person has too much time and not when all rights have been cleared and said person no longer has the time.

[Mr.Flibble]

Yup, at least if it is open, someone or some group *can* do something with it given time & effort, if it remains closed, it's dead.

[Marlon_]

Originally Posted by Mr.Flibble

Yup, at least if it is open, someone or some group *can* do something with it given time & effort, if it remains closed, it's dead.

Having the source open would also be beneficial for the likes of AROS or MorphOS as well for research and reimplementation purposes.

[Mr.Flibble]

Originally Posted by Marlon_

Having the source open would also be beneficial for the likes of AROS or MorphOS as well for research and reimplementation purposes.

Exactly - an ARM port would be most interesting!

[wXR]

I would definitely like to see the AROS guys fully able to work with/learn from the AmigaOS sources. One of my big motivations in wanting to push for the relicensing.

[FOL]

Originally Posted by idrougge

Getting the entirety of AmigaOS 3.1 to build from source is far from easy going by Olsen's accounts, especially if you intend to make a 512 kB Kickstart out of it.

Kickstarts work fine.

[idrougge]

Really? According to what I've read, graphics.library depends on the old Greenhill compiler not only because of compiler intricacies, but because other compilers can't generate as compact code.

[gulliver]

Originally Posted by idrougge

Really? According to what I've read, graphics.library depends on the old Greenhill compiler not only because of compiler intricacies, but because other compilers can't generate as compact code.

Yes, the current code depends on that specific compiler features, but after all, it is just C code.

So someone with a lot of spare time can adapt/rebuild that code for other more available compiler. It is not impossible by any means, but it will certainly take a lot of man hours.

[matthey]

Originally Posted by idrougge

Really? According to what I've read, graphics.library depends on the old Greenhill compiler not only because of compiler intricacies, but because other compilers can't generate as compact code.

It is intuition.library which is compiled with the old Green Hill's compiler. The code was just never adapted to a newer compiler. I doubt it has anything to do with compactness of code as I expect intuition.library would be much smaller (and faster) without the stubs used to convert function register arguments to stack based arguments as the compiler did not support register arguments (the standard 68k AT&T ABI specifies stack arguments only).

Originally Posted by gulliver

So someone with a lot of spare time can adapt/rebuild that code for other more available compiler. It is not impossible by any means, but it will certainly take a lot of man hours.

It shouldn't be too difficult but it is likely a monotonous and error prone job so it would require some testing afterward.