"Hyperion and Cloanto allegedly close to finalizing settlement"

[cgutjahr]

This is just Thomas voicing his dislike for open source again, and bringing up the same old FUD. No, open source does not mean "everybody does what he wants". It means Mr. Richter and his team  could work on an official AmigaOS branch - without having to team up with a company he himself accuses of pirating old versions of AmigaOS - while other people could work on CosmOS, or AROS, or MorphOS and would be largely ignored because they're not 'official'.

A good model, if you ask me, would be to hand over control to a "board", with multiple interested parties on the board, then negotiating on such board what the direction should be, then drive the process from there

That would only be a good model if the members on that board would actually pay developers - which is not going to happen (aka the Hyperion approach) or it's going to result in astronomical pricing for the OS and micro-updates you have to pay for (aka the A-EONkit approach).

AmigaOS development needs to be developer driven, obviously. Whoever controls the trademark (in an open source model) or the IP rights (in a closed source model) should get some input  of course: release cycles, desired features etc. But the idea of unpaid developers taking orders from anybody sounds like a recipe for disaster.

[cgutjahr]

Go to my github page, see how much I dislike open source. Please do not bring up claims like that. There are places where Open Source works, and there are places where it does not. It depends on what you want.

Okay, let me clarify: You don't like having your little babies open sourced. You were against open sourcing P96 - which would have been a tremendously good thing for the Amiga - now you're maintaining it. You are against open sourcing AmigaOS - and you 're the one developing it. I can see a pattern here.

AmigaOs does not have anything like that.

What does any of this have to do with AmigaOS?

Current situation: Your team develops AmigaOS 3.2. Releases it. Doesn't get paid. Then works on 3.3.

Hypothetical open source scenario: Your team develops AmigaOS 3.2. Releases it, along with a source code archive. Doesn't get paid. Then works on 3.3.

Other than that, there will be only superficial differences between both scenarios, like Cosmos starting a flame war telling people how much your coding skills suck and releasing CosmOS 3.2, and everybody laughing about him.

You're just trying to distract from the simple truth that in reality, nothing much would change by open sourcing AmigaOS.

Are you seriously mentioning "many desktops would arise" and "we would loose consistent look and feel" to tell us how bad open sourcing AmigaOS would be? Have you actually ever used AmigaOS? Do I need to make a list of available desktops, docks and taskbars for you? Or GUI toolkits and Gadtools/ASL patches?

If we go open source, we have soon a diverging operating system, an unstable platform and no solid software or hardware ground to work with.

You forgot to mention that open source AmigaOS  also abuses his wife and beats his children.

We've always had quasi branches of AmigaOS or parts of it ever since Commodore went down. People were patching their systems to death, installing (allegedly) faster serial drivers and Workbench replacements or replacing major parts of the system with backports from AROS. Guess what? We survived it. And so did AmigaOS.

It is good that we have rules how the game works, and everybody can depend on the rules.

We would still have rules. You're the only one claiming otherwise.

And If I were a developer relying on "the rules" as defined by you, I'd have started to use Reaction as a GUI toolkit when 3.5/3.9 were released. Then I would have switched back to Gadtools (?) when 3.1.4 came out, now I'd have to rewrite my projects again to switch back to Reaction.

No, I do not believe that developer driven is the right way.

So how are 3.1.4 and 3.2 developed? Who's making the decisions?

If you do it "developer driven", you create monsters like "systemd", "pulseaudio" or "cups" nobody sane in his mind is able to use.

So you're saying it would be better for AmigaOS if Ben and Timothy would call the shots instead of you?

[cgutjahr]

If you want to know who develops [p96], ask Jens. If you want to contribute, ask Jens. He's probably looking for someone to drive his product because I cannot.

So while you argued back in 2016 that P96 would end up in project management hell if it would be open sourced, it turns out there's likely no project management at all now that Jens controls it? Who could have seen that coming? It's not like we had evidence that Jens sucks at software project management, right? Oh wait, we did.

Let's keep that in mind for later.

A *lot*, because that is exactly what makes the difference. We do not have a critical mass of developers, we do not have sufficient numbers of commercial players.

The "numbers of developers" required to develop the OS doesn't depend on its proprietary or open status. Your team stays the same, so do your goals.

A lot would change, and I'm really stunned that you don't see that. We would not have one version of AmigaOs, but multiple.

Again: no we wouldn't. We would have one AmigaOS developed by one team. And maybe a fork or two using a different name, trying out different things. More likely just forks of individual components/libraries, because building Kickstart or entire distributions is a lot of work probably not all that simple.

Now let's compare that to the current situation: We have AmigaOS 3.9, Amiga OS 3.1.4, AmigaOS 3.X, AfA OS 4, AROS 68k and "BoingBag 4". Plus a whole bunch of what I refer to as 'Workbench distributions', like Amikit, AKReal, Coffin and UltimateWB. There's an icon.library replacement and an updated graphics (?) library on Aminet and Cosmos must have 'optimized' every single part of Kickstart by now. And that's just the 68k side of things...

Where's the actual change you keep warning us about?

To be able to keep development possible, we need to cut *down* the complexity, not increase it.

Obviously, the official OS sets the standards - you're not responsible for forks. Complexity problem solved. You're not trying to stay compatible with Peter's icon library, Cosmos' graphics optimizations or the non-standard init process of the Vampire right now, so why pretend you'd have to start doing that once AmigaOS is open source?

Which rules? Once it is open source, the ghost is out of the bottle.

Technically, it already is open source. I bet I can find the source on the net in less than a minute. And nobody's taking down projects like BB4 or Coffin (while, ironically, *your* project's right to exist is questioned in a court battle).

The rules are set by the party controlling the trademark. It really is that simple. Once that party ceases to exist or does what all of its predecessors did - i.e., fail miserably, creating tons of damage in the process - we might get the anarchy you're so afraid of. But we had that anarchy happen even without an open source OS, so once more: not different at all.

I would say that AmigaOs would be better if we had a specialist in UI design (and not a lawer) who would be central in decision making, yes.

A specialist in UI design is a developer of course. I was using the term as a shorthand for "people who work on the OS" - that includes a UI interface designer just like it includes the beta test coordinator or the translations/documentation people.

What we would need is a management that believes in its product (and not its money it may or may not make).

You're doing what you always do in these discussions: You admit the status quo is not good at all, but then go on to claim that a proprietary development path at least has the potential to improve things - while an open source approach will definitely kill and destroy everything and has zero chance of success.

So let me get this straight: We should trust you, that Ben doesn't completely f?ck up your OS project, like he f?cked up OS4 - half of which is now owned by another party that hates Ben's guts? And the reason we should put this much trust into him is that open source will definitely kill the Amiga?

My head hurts.

[cgutjahr]

MUI vs ReAction, NSD vs whatever the other thing was, P96 vs CGX, PowerUp vs WarpOS, and, of course, the big one, OS3 vs OS4 vs MorphOS vs AROS.

• All of those were created despite AmigaOS being closed source
• All of those were created while there was no active development of the official OS
• All of those were resolved the second the official rights holder declared one of them to be the official solution

Why is critical mass important? Look at what happened with AWeb. We had an open-source web browser!

We also had an open source DPaint! Look at what happened! What do mean, "it was totally outdated by the time its source was released"?

As much as we want it to be true, open source isn’t a panacea for every situation.

Nobody's claiming that. But the situation is really, really fucked up. And Thomas and you are arguing "let's continue to try what we tried countless times in the last 25 years, one of these days it has to work, right?".

Instead of one slow-moving branch we could have 4 dead branches.

Why? Because Thomas and his guys rage-quit when the sources are released?

Thomas is also right that a developer driven model is rarely one for long-term success.

"Long term success"? We're discussing AmigaOS 3.x - where do you want it to go in 2020?

[cgutjahr]

Plus a lot of trolling, plus a lot of "here is my branch of library YZX, take it or leave it", plus a lot of complains "waaa, they didn't take my patch". Thanks, but I guess we do better without that in a more ordered process.
[...]
Yeah, right. Except for the whining. "Waaa, your Os is broken, it does not work with software A". "But that requires the fork from B, not our base" "Waa, you suck".

So, when you say "anarchy" you actually mean people not being nice to you?

We *must* cut complexity down

You're not talking about complexity, you're talking about variety. Are you arguing we need to kill AROS 68k and Cloanto's distributions?

There is a consistency from 3.1 to 3.5 to 3.9 to 3.1.4

You're dodging the actual question. Let's ignore the lack of Reaction and say 3.9 and 3.1.4 are the same - that still leaves Amiga OS 3.1.4, AmigaOS 3.X, AfA OS 4, AROS 68k and BoingBag 4 plus the individual patches and libraries released on the Net. The existence of all these branches did not stop you from releasing 3.1.4, did it? So where's the limit before we loose you? Five, six, seven branches?

And you do what you always do: Pretending that open sourcing the Os would solve these problems.

I'm saying it at least gives us a chance of fixing these problems, while what you are doing right now just feeds the lawyers and is guaranteed to create tons more problems when (not if) Hyperion collapses.

I'm not "pro Ben", I'm "pro controlled development, no matter who does it".

Right now, Ben is controlling development - and you prefer that to open source. So as far as this discussion is concerned you are pro Ben. Which means you should be able to explain why Ben's control over your project doesn't mean it's doomed.

[cgutjahr]

Cloanto, however

For somebody who's constantly complaining about trolling you sure love to comment on other people's intention, work or competence. Any mud fests with Gunnar recently? Or public declarations that Cloanto is distributing illegal versions of P96?

Not sure why you keep referring to Cloanto, I'm just discussing a general concept.

AfA 4.x I consider a dead horse

The latest AfA release is younger than your latest release, IIRC. That plus AROS 68k leaves two branches, yes? Plus 3.X from Cloanto. So you're fine with two or three  branches? How many is too much?

No, the world is not black and white. Meaning I'm against open source does not mean I'm for Ben, that is just naive.

You're creating valueable IP for Ben that can be held hostage, just like OS4 has been held hostage for a decade. You're creating revenue for Ben. Given the choice to open source or stay with Ben, you'd stay with Ben. Hence, in the context of this discussion you're pro Ben.

Explain to me how giving 3.1.4 for free to Hyperion makes it more likely that some sort of comittee of skilled benefactors will get a say in what's going to happen? Because it worked so well for OS4?

[cgutjahr]

The latest AfA release is from 2016. The latest AmigaOS update is from 2019.

I was talking about the follow up project, I think it's called "AfA One". That one had a release in October.

[cgutjahr]

But I would argue that all of those were created *because* we'd lost the governance that Commodore provided in its stewardship of the OS.

I don't think 'governance' is the right term here. This is/was a commercial market, and both hardware and software developers would fill whatever niche they either saw as commercially viable or bothered them in their personal use of the Amiga. You can't just forbid that from happening.

And even though one or the other has been declared the "official" solution, the situation is far from resolved--MUI is still used despite ReAction being the "official" choice for OS3/OS4. If we still had strong governance a solution would have been found before a split happened, and in a way that would have not left proponents of an alternative feeling personally offended.

The GUI is actually a good example. 1.x had no proper GUI toolkit, which is why everybody created his own. Then Gadtools came along, but nobody used it because a very large chunk of users was still on 1.x. Once the majority of users were able to run Gadtools applications, the system was already outdated again, which is why we got all those Gadtools extensions or completely new GUI toolkits.

Neither Commodore's presence, nor the closed source nature of AmigaOS nor CBM's announcement you should use Gadtools from now on did do anything about this situation. What finally solved it was continued OS development. MUI still being alive is a special case, it simply survived because the competition adapted it as their solution.

No, because the 4 branches will each start out with passionate defenders who ultimately aren't numerous enough to sustain them in the long term. Meanwhile everyone else will be confused as to which one to back and in all likelihood will end up backing none of them and just lose interest and walk away. That doesn't benefit anyone.

Nah. Let's say a hypothetical open source  3.2 comes out, developed by a team lead by two respected and well-known developers, one of which has been at this for close to three decades, and it's delivered in a big shiny Box labeled "AMIGAOS 3.2". There will be absolutely no confusion as to which project one should support. If history tells us anything, there will be threads where people announce they're going to buy it, despite not having a need right now - just "to support the Amiga".

MorphOS had a head start, ran on much, much better and cheaper hardware that was continously available - and in contrast to OS4, it has seen tons of development in the last decade. But as soon as Bill Buck stopped throwing suitcases of money (or at least promises of such suitcases) at the whole thing, it got pretty much irrelevant over night. I could say similar things about AROS, but OlafS is here, and that always gets him worked up - so I won't

My ultimate interest is preventing this from happening again:

I get that. But I don't think it's much of a problem in 2020 - partly due to AmigaOS now being a retro hobby toy, partly due to the low number of skilled developers. If I have two or three of those on my team, I control 80% of the qualified work force

And let's not forget that Hyperion/Ben is the party that allegedly torpedoed the very early attempt to join forces and create one single PPC based AmigaOS from components controlled by Amiga, H&P, Hyperion and the MorphOS team (*)  and established a 3.x branch after Cloanto did - one who's legality is currently questioned in court - and is hated by a very large chunk of the developers and managed to get their proprietary PPC OS which was completely under their control split into two efforts. "Unity, peace and common standards" is not the first thing I think about when hearing "Hyperion".

Unfortunately, the choice is not "open source or some hypothetical well-meaning entity lead by Jay Miner's grandson" - it's "open source or Hyperion".

(*) not saying Hyperion was the only guilty party

[cgutjahr]

Err, what? You were not on board, so you can hardly know *when* 3.1.4 was established, but it was quite a bit earlier than you may believe.

Cloanto's first (digital) 3.X releases are from 2004. I'm not sure when the first physical releases happened, probably a decade later.

[cgutjahr]

Fair point, developers will fill gaps the vendor has left open. But originally all those toolkits were to support products at the application level. Now we have a problem of competing components at the OS level [...] now are far beyond GUI toolkits. Now it seems like we've got multiple iterations of icon.library, workbench.library, exec.library, scsi.device, ixemul.library, etc.--

You can either argue that open source will create a mess, or that we already have a mess - not both

ixemul.library is evil, because many ports require it and it was always a bitch to pick the right version and set it up correctly. It has nothing to do with the OS though.

The others are not really an issue - if you're using one of those patched execs (which  have existed for decades, I think 'Executive' showed up in the mid-nineties?), you're on your own if something goes wrong. Most people never touch them anyway.

variants that all have the same name. It's getting too difficult to keep track of them all and they're not all interoperable.

That's the point of ongoing development: You don't need to keep track of anything, you just purchase/install the next OS update if you feel like it. This whole "hunting down libraries with a version number bigger than what I have installed" thing (aka "BoingBag 15") has become some sort of fetish for a select few, the vast majority would be more than happy with bi-annual OS updates.

Well, that's what we'd hope would happen and what logically *should* happen under a (strong governance!) open-source model

It is what is guaranteed to happen. "Open source" just means that for each release of the OS you make, you also release the source. The rest of your development process can stay exactly the same. Thomas, Olsen and the others decide what goes into next release of the official OS, and what doesn't, then implement that, then make their release.

They have THE NAME, the reputation, the experience and the advantage of having familiarized themselves with the code and build process for years. The idea that I could just take their Workbench source, compile it and suddenly become the next big player in this market thus forcing Thomas to rage-quit is nothing but fear-mongering, sorry.

But that's kind of my point, right? There's been a split and there's no going back from it.

I disagree with that, but I don't think it actually matters. Split or no split, we were discussing the future of AmigaOS development. That a split happened while it was closed source, is more of an argument for going open source, no? In an open source world, I can take whatever useful stuff is implemented in any of the forks and bring it back into my source tree. In a proprietary world, I can't...

Sigh. Yeah, that's all true. But isn't even open source is a fantasy? Hyperion's not going to open-source it, and neither will Cloanto if they prevail, right?

Battilana has been stating publicly for years that he "wants" to open source it. I'm not going to just accept that as a given if he prevails - I've been watching this soap opera for decades, after all - but at least he's consistently making the same claim.

[cgutjahr]

3. We don't currently have the systems and structures in place for successful open-source governance.

Those are just big words, what do they mean? I explained why (and how) AmigaOS development could continue exactly like it did IMHO - what are those "systems and structures  for open source governance" we allegedly need and why do we need them?

[cgutjahr]

I've been thinking about this sort of thing for a few days. this 2006 article from Sun Microsystems is a little old but has some good points, getting into the merits and pitfalls of open source. Bear in mind that he's speaking about large, corporate-backed open-source efforts; we'd need something appropriate to the scale of our tiny community:

Sun is actually a very good example for what I'm saying. The reality is that Sun, despite their promises and announcements earlier on, kept developing OpenOffice pretty much like a closed source project: The original development team in Hamburg was doing most of the work - and while it was theoretically possible for external developers to submit patches, those patches very rarely made it into the actual source code.

The situation got so bad that members of the Linux and free software communities established a fork called go-oo to make sure they could at least customize the build process to their needs - all versions of OpenOffice for Linux/BSD were actually go-oo builds, not OpenOffice builds.

But as you're well aware, none of that stopped OpenOffice from being a massive success - without any "open source governance" to speak of (other than for managing communities, translation efforts etc., obviously).

Fundamentally I'd say it's a combination of institutional management, project management, and community management.

If you choose to go that way, yes. But nobody is forcing you to do that.

I don't think you replied to my suggestion of not changing the development and release process at all other than releasing a source archive with every binary release you make?

Needs zero additional governance, and no project or community management skills that aren't already required anyway. Yet, it guarantees we're not facing the next dead-end if Thomas leaves the Amiga scene for another 15 year sabbatical or Hyperion turns out to be not as stellar an outfit as we all assumed (or simply vanishes).

The professionalism and discipline cited above are key. We need (1) user- and developer-backed decision makers to set long-term goals and priorities of individual releases, we need (2) community liaisons to stay on the pulse of user needs and to be the first line of communication so that (3) skilled developers can focus on executing the roadmap for a given release.

No we don't. All of that would be nice to have - for any open or closed source project. But we don't have them right now, and we're getting by somehow. We'd still be getting by without them if the project code was freely available.

The professionalism and discipline are essential in the process so that everyone's opinions are respected, even if some people's opinions aren't adopted verbatim. That's how we prevent (further) fragmentation.

Are you saying Thomas is not professional enough to prevent fragmentation? That's mean. Or, to use his words: "stop trolling, you have no clue"

We established that fragmentation has become pretty bad already. We also established that the source code is already easily available and that people are doing things with it. We also established that people don't care much about copyright anymore - that includes the 3.1.4 team, which works for a man Thomas himself accused of pirating 1.3. Yet, despite all of this, 3.1.4 was quite a success. I don't see any evidence things might get worse on the fragmentation front.

Can you name any coder who could pull of a AmigaOS/Kickstart fork and isn't (a) long gone, (b) busy with AROS/MorphOS/OS4/ExecNG (c) stupid enough try to compete with a competent, well-respected team that's been working on the code for years (d) disillusioned by the prospect of being ignored due to not having THE NAME (e) laughed out of town every time he suggests he could improve some library? I can't.

[cgutjahr]

Yes. Me. (-: Let's consider: Again, for the sake of the argument, why should I be stupid enough to work for Cloanto for nothing if I can work for myself for nothing? (-;

For the same reason you're stupid enough (your words, not mine) to work for Hyperion for free?

And again: I have no idea why you keep bringing up Cloanto. It's almost as if something regarding open source came up in the recent settlement discussions and you feel threatened by it. Care to comment?

I'm discussing the general idea of freeing the IP, not some Cloanto takeover. If Ben wants to free the IP, or Ben finally sells the IP to Trevor and he's freeing it - I'm all for it.

[cgutjahr]

Open Source can be a commercial success, but that requires a business model. I do not quite see that - if you get can the thing for free for download, well, as far as I know the Amiga users, there would be no point buying anything. So why support an open source distribution company?

I'm just guessing here, but I'd say Cloanto's business model is not selling operating systems, but selling user-friendly, all-in-one, easy to install nostalgia trips (i.e. Amiga/Commodore Forever). Cloanto's business does not depend on OS sales. He's been selling a bundle containing all Amiga ROMs for something like 70 cent on Google's playstore for ages, I doubt he's getting rich from that.

Why buy AmigaOS if it's open source? Because...

• ...only the commcerical version would have official branding
• ...collectors buy (a) anything in a Box and (b) anything with an 'official' Amiga mark
• ...ROMs and disks can't be downloaded from the net
• ...people want to support the effort

Should be enough to support the effort of making a physical edition.

Btw.: I admire your guts talking about 'business models' while working for/with Hyperion.

[cgutjahr]

This discussion is pointless. Somebody working for Hyperion critisizing the lack of a viable business model for an AmigaOS open source project? An AmigaOS developer pointing out that Linux is "not successful" on the desktop? An AmigaOS 3 developer worried about "loosing consistent look and feel" when we go open source? This is a parody of a discussion, at best.

Thomas is simply coughing up whatever comes to his mind first and sounds scary enough - just to completely drop the issue once he gets to hear some counter-arguments. Apparently, open source would be terrible, because "so many different desktops would arise" - until you point out that we never suffered from a lack of desktops. Then it turns out desktops do not actually  bother him - only Kickstart forks do. But obviously, none of the existing Kickstart forks count as counter examples, because... reasons. And did you know THERE ARE NO PROFESSIONALS WORKING ON DESKTOP LINUX?

I think there are two things we can take away from this 'discussion':

1. Thomas will walk away if AmigaOS becomes free software. So there.
2. Stop trolling, you guys.