AmigaOS 3.1.4 withdrawn from market after Amiga

[Thomas Richter]

Originally Posted by kolla

The "mess" I talk about is all the incarnations of Amiga shell you can expext to find on 68k Amiga systems. Someone running "OS 3.9" can be running anything from v40 to v46, depending on what kickstart and ROM updates the person has bothered/been capable of installing. Writing scripts that cover all, means dropping all "modern" features and working around various bugs that may exist

Ah, so the "updating" part is now not the problem? So what do you actually expect? No one working on the Os, which means "no changes at all"?

Strange that *exactly you* bring the argument of things that things should remain untouched. Isn't change exactly what happens on a daily basis in OSS systems?

Now I dare to provide a new built-in shell feature such as pipes, and you make a ramble about it because you can no longer use an undocumented(!) inofficial(!) feature of Os 3.1 such as the PIPE command and two shell variables?

I am sorry. I don't get it.

[Thomas Richter]

Originally Posted by Gorf

Guess what would have prevented this kind of situation .... guess what would prevent it from happening again ... why not learning from the past?

Making proper contracts would have helped for a start, indeed.

In this respect, Os 3.9 was... considerably weird and badly thought about. It was a collection of third party contributions which were licensed for a limited time to H&P for distribution (basically, the contract was >20 pages IIRC). In particular, contracts were even individually negotiated. While I do not know Joane's contract, mine included that my tools fell back to me after 2 years.

While good for me (and no harm done in my case), this was a particular problem for other contributions.

So what can we learn from this: If you want to build an operating system, better make sure that you have a long-term strategy for its development.

Unlike you, I do not believe that OSS is the only answer for such a strategy.

[Gorf]

Originally Posted by Thomas Richter

But you are aware that I'm not making the rules, correct?

You are not making the rules, but you are following them.

I can not tell how big your influence in this matter actually is, but as you and your coworkers are the ones making it happen, you are certainly not entirely powerless either.

By following the mentioned rules, you are also embracing them and enabling a company like Hyperion ... a decision I personally would not have made.

(enough from my side now - let's see what happens in the future)

[Minuous]

@Gorf:

Out of the two options they are much better than CloneToo.

[Gorf]

Originally Posted by Thomas Richter

Unlike you, I do not believe that OSS is the only answer for such a strategy.

No I do not think that - not in general.
But I am strongly convinced it is the only viably solution in our case.

[Gorf]

Originally Posted by Minuous

@Gorf:

Out of the two options they are much better than CloneToo, surely!?

the devil and the beelzebub.
So the correct answer must be none of them.

[Thomas Richter]

Originally Posted by Gorf

You are not making the rules, but you are following them.

Ok. Alternative: I do nothing, and instead complain, like everyone else?

Originally Posted by Gorf

I can not tell how big your influence in this matter actually is, but as you and your coworkers are the ones making it happen, you are certainly not entirely powerless either.

No, certainly not. But you should understand two aspects here: There is a legal aspect, onto which I have precisely *zero* influence at all. I am unhappy about it, and I am not particularly happy about Hyperion. I am not particularly happy about Cloanto either.

A considerably better solution to this would be to solve this market wise: Cloanto develops, Hyperion develops, and whoever has the better product makes the money. Works fine for me.

Then, there is the technical aspect: It is quite understandable that those people that contribute more have a better standing of the technical development. Not the only one, of course. If there is nonsense, and co-workers or beta-testers report that a feature is not to their liking, it is changed.

There were no comments on the pipes, for example. The current implementation is considerably simpler and easier than the "detour" through an external command, and it is - for the future - also easier to maintain. As no negative comments came in, the feature stuck. Quite simple.

So even beta-testers can make a difference by simply rejecting a feature, you do not need to program, even.

Originally Posted by Gorf

By following the mentioned rules, you are also embracing them and enabling a company like Hyperion ... a decision I personally would not have made.

This is a hard decision to make. In a situation like we have here, one can do two things: Either accept it, and be quiet about it as - without having a hand on the development - one cannot do anything. I stayed quiet for a long time on Os4 development and pulled myself out. Simply because I had the feeling that the whole direction was wrong. But that did not mean that kept shouting every five minutes how bad the thing was. Which, quite frankly, I do not even know. I do not care, I just consider the whole Os 4 thing a silly idea.

But then, you can also take the chance and try to make a change. There was a window of opportunity.

Keep silent forever, and let Amiga rot - or get something going, well knowing that you can make a difference and drive the whole process in the direction you consider the right one, well knowing that you don't have the "dream parter" on your side?

Now, this is a hard decision to make - what would you have done?

[Gorf]

Isn't change exactly what happens on a daily basis in OSS systems?

Actually no - changes that would jeopardize compatibility are generally discarded - it is very hard to deprecate something in OSS. At least much harder than in any closed source project.

[Thomas Richter]

Originally Posted by Gorf

the devil and the beelzebub.
So the correct answer must be none of them.

Which would have improved the situation ... how exactly?

Look, if you truely believe that OSS is the right answer... Contribute to AROS. That is also perfectly fine for me. I'm "burned" because I've seen the Os sources too much from the inside.

My two cents on this, despite the all-legal problems that may arise if my hands touch this source: If I would write an operating system (OSS or not), I would certainly not want to repeat the design problems AmigaOs had. The whole Tripos/BCPL/C integration is ... complicated. The lack of resource management and process isolation is just... silly today.

If I would want to say something in direction of AROS, then that: Please write a proper operating system. Not the kind of nonsense AmigaOs piled up over the years. As it does not need to stay compatible, one could certainly do better than anything AmigaOs could ever become.

[Gorf]

Now, this is a hard decision to make - what would you have done?
---
Which would have improved the situation ... how exactly?

With your skillset and expertise?

Most likely I would have tried to improve AROS and the 68K Kickstart replacement.
I would imagine the two years of work, the 3.1.4 team did spend, would have made a tremendous difference to AROS in all regards.

The second most likely scenario: I would have insisted on some OSS clause in the contract, that becomes effective after a while or in case of bankruptcy.
(well - maybe you did )

[Thomas Richter]

Originally Posted by Gorf

Actually no - changes that would jeopardize compatibility are generally discarded - it is very hard to deprecate something in OSS. At least much harder than in any closed source project.

Ah, then how come that everytime I update linux, old programs no longer work? Yes, the interface between the kernel and user space is robust and static. That I can agree upon. But the user space itself is not. Changes are made without good justification, and old interfaces are put aware and replaced by something that is "almost, but not quite the same". I believe I wrote like two or three versions of an Audacious/XMMS plugin, because they kept changing the thing - for no reason. I finally gave up. Installed the old thing, with all its dependencies, and let the issue rest. Also, the kernel space kept moving. There are no set interfaces between kernel modules. I have here an old IBM laptop, where - with lots of effort - I was able to get the intel graphics to work. No longer in the next kernel, something changed, the thing broke. I have here an installation of a VM, which I need to recompile for every new kernel update, because some interface in the kernel networking infrastructure keeps changing. Parameters are omitted or added - for no apparent reason. The old functions could have just stayed in, or default parameters could have been added to keep the interface stable. Of course, that is not "very satisfying" from a developer point of view, and they are not part of the "user space" interface. The problem is here the lack of a properly defined interface between kernel modules. There is nothing like the "scsi.device" or the "graphics.library" where you have a documented API you can use that does not go away next day. The only reason why the monolithic linux kernel does not break under its weight is because there are so many people that keep working on it. A company would not be able to keep a code with so little modularity updated and running. AmigaOs does not get many things right, but at least one thing is done correctly: Modularity. It is a quite tiny microkernel. Of course, with many important features lacking, given its age.

[Gorf]

Originally Posted by Thomas Richter

If I would want to say something in direction of AROS, then that: Please write a proper operating system. Not the kind of nonsense AmigaOs piled up over the years. As it does not need to stay compatible, one could certainly do better than anything AmigaOs could ever become.

OK - that would be the third option

Fuchsia is quite interesting (kernel-wise).

[Thomas Richter]

Originally Posted by Gorf

I would imagine the two years of work, the 3.1.4 team did spend, would have made a tremendous difference to AROS in all regards.

This could have had bad consequences for AROS, however. Likely outcome: One of the two would have detected that some AROS code looks particularly familiar, and presents a suitcase with "proves" that the code was "stolen" from AmigaOs.

Believe me, it is really better for AROS if I do not look, do not touch. That is one of the better insurances that nobody from the two comes along and makes some claims on it.

Originally Posted by Gorf

The second most likely scenario: I would have insisted on some OSS clause in the contract, that becomes effective after a while or in case of bankruptcy.

That sounds legally complicated. In case of bankruptcy, the rights do not go away. In fact, all the assets of the company need to be used to satisfy any open claims or demands. In case of Hyperion, it gets even more complicated because they do not own the whole thing. They have an exclusive development license (as far as I can read it), so they can hardly transfer any rights to a third party they do not have.

[Minuous]

@Gorf:

IIRC one of the only restrictions on Hyperion in the Settlement Agreement is that they are not allowed to release the source code. So there is no point hassling ThoR about open sourcing it at this time. Those wanting to work on it in the meantime would be better to join the official team. I'm willing to Reactionize a couple of Preferences editors, at least, as part of a Hyperion AmigaOS 3.91 (or similar) release.

[Gorf]

True: Linux is a big mess in many places
That is why all driver work I did so far, I did on BSD -- much much cleaner structure and code.

Your problem with updates seems more distribution-related - as you admitted: The Linux-Kernel interface is very stable.
I can not really support your observations: after a recompilation usually everything is fine.
(Used to BSD "ports" - Gentoo provides a Linux-equivalent)

Still: I am by no means a fan of monolithic kernels or even posix.

And yes: I do like how AmigaOS approaches things and it would be interesting to see that translated into a modern design....

[wawa]

Originally Posted by Thomas Richter

Look, if you truely believe that OSS is the right answer... Contribute to AROS. That is also perfectly fine for me. I'm "burned" because I've seen the Os sources too much from the inside.

thats exactly how this exclusion works. even if you yourself have probably a well established bias against this approach, it doesnt mean it doesnt affect others, who might have been considering it. im though thankful for your educated opinions and "white room" kind of support or advices passed to aros team so far.

edit: i meant "clean room" of course. donno why that opposition reminiscence..

[malko]

Originally Posted by SpeedGeek

[...] My dictionary lists "Unclear" as one of the words which define "Dubious". Perhaps your understanding of English is not so good. [...]

Humm... two sentences and already a personal attack. It promises good arguments from you .

Originally Posted by SpeedGeek

[...] You can ignore my (jurors) opinion, and the lawyer's representing the Amiga parties and Cloanto but what about Hyperion's own press release? [...]

I really don't see what you want to show with a news taken from Hyperion's site. As a "juror" (how you call yourself) do you consider that a news taken from a commercial site has any legal value ?
I have heard that US right is different from CH right but... is it really that much ?

What kind of legal implication do you want to link to a "marketing announcement" posted on a commercial site as an announcement's sole purpose is to inspire the defined audience the wish to learn more and to get engaged with the product you want to sell.
Furthermore, what from this 2009 announcement do you consider relevant enough that can still apply in 2019 ? It's not a contract, it's not an signed agreement. It's a news posted in a commercial site.

Are there any customers who have been fooled by the 3.1.4 OS release regarding this 10 years old news ?

Just to mention two example, are you going to sue your favourite soda brand because it's refreshing only because you took it out of the fridge or merged it with ice-cubes ? or your favourite fast-food because what is in your tray is far from being identical with the nice and attractive picture that is behind the cash counter ?

Again, if US right is different, you are welcome to let me know.

Look at this one and tell me from which company it is (I have replaced a name by "xxx" to challenge you) :

More than three decades of uninterrupted ease of use, power, beauty and excellence: "xxx" closes the circle between gaming, productivity and preservation of digital culture while adding new features and providing access to a universe of free and legal downloads. [...]

BEAUTY, EXCELLENCE, EASE OF USE, POWER, PRODUCTIVITY, NEW FEATURES, UNIVERSE... whaaaaa...
So, what kind of legal implication do you want to link to that ? It's an announcement, it's marketing. No difference with the news you quoted.

Regarding marketing announcement, I quite like this citation which is attributed to a former French president :

Originally Posted by Jacques Chirac

Promises are only binding on those who receive them.

So let's wait the jurors' decision (in their soul and conscience).

[apex]

Nice, finally I found the thread with the most Amiga trolls!

[grond]

Originally Posted by Thomas Richter

Hold on. This is not about *my* control. Want to contribute? No problem, come on board.

To do that I would have to sign a NDA and tie myself to Hyperion. I'm not going to do that. So, yes, your control, because you did it.

No problem either, write something nice, put it up on Aminet. If you believe you can create software without bugs, it also works with me. Demonstrate!

It's not about producing something without bugs although having produced that one obviously hurts you a lot. It's about being able to fix other people's bugs. And I have demonstrated that I have done that before. The point still is: an open-source development model would be better than the closed-source one.

I have issues for "rants", for complaints for the point of complaining. You see such people a lot. People that do nothing, contribute nothing - in no particular way.

Yes. People can rant about commercial products as much as they want, no? If it were open source, you could just say: go fix it yourself if you think you can do it better. But expecting people to stop critisising your work, sign an NDA and contribute to a commercial product for free is a bit too much. You put yourself into a position where you are found responsible while others cash in on your work. Must be a very comfortable situation for Hyperion selling bugged software. I mean, even better for them, now they can sell the product and then the bugfixed update (which will, of course, come with other new fixes and features so nobody can say they make people pay for the bugfix of a bug that never should have happened).

I haven't heard a single word from "specific persons" saying, "oh how nice that we can now boot from large disk drives, but I found this particular problem...".

I do appreciate your work. What I don't like is your attitude about how open-sourcing AmigaOS would be worse than the current situation.

What I am not happy about are people that scan for every little bug, and then make a rumble about every little bug just to drive people away.

Criticism can be both positive or negative. If positive, it is an excellent contribution. If negative, it helps no-one.

Of course, your constant out-of-context-ranting about the Vampire's lacking MMU, construction of how many unknown bad side-effect the reduced FPU precision of the Vampire might have, how your i5 is so much faster than a Vampire and so on was entirely positive criticism, right?

[grond]

Originally Posted by Thomas Richter

Ok. Alternative: I do nothing, and instead complain, like everyone else?

Actually you come across very much as a complainer. You have been complaining about complainers for several pages.