@Karlos

I think selecting Linux was a great choice. Linux has been battle tested in the stock market for years and runs most if not all Wall Street servers. A bare metal based Amiga Linux distro like Amithlon would be awesome today if it had not been abandoned. Haage & Partner really screwed this one up and left the Amiga scene soon after. A missed opportunity for the developer of this great software unfortunately. I sometimes wonder if McBill helped screw this one up too.

_________________
Sent from my Quantum Computer.

@Karlos

Quote:

Karlos wrote: @agami I dunno, it wasn't aimed at you, there are many people that seem to cling to IP and trademarks keen to stifle anything that looks like progress. Why, do you think it applies?

Quote:

Just reading Bernie's posts here makes me appreciate what we truly lost with Umilator and Amithlon

Because when I read Bernie’s posts, and your enthusiasm, I’m nostalgic for what can still be.

While I don’t think the sources themselves have direct monetary value, though I’d happily compensate Bernie if need be, I have never shied away from conversations of commercial nature on this forum site.

I’d like to see a new high performance JIT 68k Amiga out there, and for it to have any level of staying power, there would have to be some commercial components to the offering. Not to make anyone rich, but to stave off oblivion.

Commercial talk amongst techgasm geeks can often make one be perceived as a vulture. I know that’s how the Apollo Team considered investors in the early FPGA days, and I didn’t begrudge them not entertaining my interest in helping to fund them.

_________________
All the way, with 68k

@Karlos

Quote:

Karlos wrote: Whether Linux was the right one or not is a different question.

That was exactly the question I raised (if it wasn't obvious).

Amithlon is stuck in time, with a Linux 2.4 kernel and all the limitations that comes with that. Umilator may have been maintained to perhaps some 3.x kernel or so, and then what? The point is, a closed source projects that depends on "hacking" a rapidly moving open source kernel is not a sustainable, and in case of the Linux kernel... too much of a legal mine field. Solution to this problem is either to go open source and license compatible from the get-go, so that transfer of knowledge and maintenance can be spread among more heads, or chose something else than Linux.. doesn't need to be baremetal, there are plenty of alternatives to the Linux kernel.

Emu68 is in a different position entirely, its biggest "problem" is exactly that it is (for now) tied to certain Raspberry Pi models specifically. But does it run on, for example, a virtually "raspberry pi" via Qemu on Apple Silicon, using either macos' hypervisor, or any other aarch64 with KVM under Linux?

Last edited by kolla on 01-Mar-2023 at 03:47 AM.

_________________
B5D6A1D019D5D45BCC56F4782AC220D8B3E2A6CC

@DiscreetFX

Quote:

DiscreetFX wrote: A bare metal based Amiga Linux distro like Amithlon would be awesome today if it had not been abandoned. Haage & Partner really screwed this one up and left the Amiga scene soon after. A missed opportunity for the developer of this great software unfortunately. I sometimes wonder if McBill helped screw this one up too.

You describe here exactly why such a project was doomed from the start.

Amithlon/Umilator would have been awesome and survived a heck lot longer had it been open source from the start and not been dragged into the entanglement that is “commercial amiga industries”.

Imagine if Hyperion had convinced Michal to develop Emu68 exclusively for them, to be licensed and sold with OS 3.2+ … ew much - abandon ship!

Last edited by kolla on 01-Mar-2023 at 03:47 AM.

_________________
B5D6A1D019D5D45BCC56F4782AC220D8B3E2A6CC

@DiscreetFX

Quote:

I think selecting Linux was a great choice. Linux has been battle tested in the stock market for years and runs most if not all Wall Street servers

Yes, where it is used to run “bare metal” emulators of DEC VAX, Alpha and Intel Itanium systems so that the stock market software that only exist for OpenVMS can continue to dominate - the parallel to Amiga desktop is striking!

_________________
B5D6A1D019D5D45BCC56F4782AC220D8B3E2A6CC

@kolla

I can't speak for Bernie but if I had to guess, Linux was chosen because it got the job done and allowed him to realise his ideas. Almost every developer I know doesn't spend long agonizing over the possible legal and long term maintenance aspects of their technology choices when working on something.

If we did, we'd never start anything.

Quote:

The point is, a closed source projects that depends on "hacking" a rapidly moving open source kernel is not a sustainable, and in case of the Linux kernel... too much of a legal mine field. Solution to this problem is either to go open source and license compatible from the get-go, so that transfer of knowledge and maintenance can be spread among more heads, or chose something else than Linux.. doesn't need to be baremetal, there are plenty of alternatives to the Linux kernel.

The only real obstacle here is the fact the the target is closed source. Everything else is engineering. I still think Linux would make a good choice today, if for no other reason than it has a wide selection of up to date drivers for modern hardware, especially on the display side. Are there better kernels? Maybe. Do they have the same level of hardware support?

Last edited by Karlos on 01-Mar-2023 at 09:40 AM.

_________________
Doing stupid things for fun...

@Karlos

I can't speak for Bernie either. But, back then he was a linux guy, and had just recently released UAE-JIT, so i think it made sence for him to use linux, as he was used to that environment.

_________________

@kolla

Not in my experience, when I worked at the New York Stock Exchange they were not emulating another OS or system. Native Linux applications all the way.

_________________
Sent from my Quantum Computer.

@Karlos

Quote:

With respect, I disagree.

I shall take that as a compliment.

However, given that I know what you are complimenting me on, and you don't, I equally respectfully disagree right back :)

This was never a well-designed and well-planned piece of code. Not a lot of planning is even possible, when most of the most interesting problems only become apparent after you have solved a bunch of other, quite complex problems.

In my experience, software that is organically grown like that can go one of two ways. Either it turns into a convoluted mess, or everything clicks into place, later problems illustrate the right approach to the earlier ones, and all the complexity can suddenly be addressed with some small and elegant piece of code which doesn't need any exceptions for corner cases, and which, simply by being right, often ends up solving problems before one has even realised them. Hitting on one of those is a rare but very rewarding experience.

Amithlon has some of each --- some gems I'll happily reminisce about, and some mess that I'd probably despair over if looking at it again. But the thing is, even the elegant bits (or maybe even especially those bits) have very little to teach. The magic of those bits is in the journey, in all the wrong or clumsy approaches that preceded them, in understanding what those approaches lacked that the elegant one has, and maybe in developing an intuition for how to get from there to here. And none of that is in the source --- source only tells you what, it doesn't tell you the much more interesting why.[1]
On a related note, sources don't tell you what could have been, but for the constraints of the time; And the flip side, what assumptions are needed for code to work, which may only be true in the code's specific context.

An analogy that comes to mind is it's one thing to see the famous 0x5F3759DF magic in code; It's a completely different, and much more helpful, thing to read why it works, what might break it, and why it might or might not be a good approach away from x86 Quake.


[1]: Yes, that's what comments are for. Have a guess how many comments do you think that bloody stupid 30 year old PhD student with zero commercial experience sprinkled through that never-to-be-seen-by-another-soul late night project.

@DiscreetFX

Quote:

DiscreetFX wrote: I don’t want to and am not trying to, he’s free to chime in of course.

I thought that's what I had been doing?

@Karlos

Quote:

I don't know how practical a bare metal solution like Emu68K would've been for that era of PC hardware.

Not at all practical. Still isn't. Nobody does bare metal PC for actual software, because you never know what your (user's) metal will be.

If I wanted to do bare metal, then I'd probably either also go with the Pi[1], or, for more fun and also more pain, try my hand on the Apple Silicon laptops, which are updated somewhat infrequently, and for which Asahi Linux could serve as a reference.

And yes, sometimes I still feel the itch. But then I spend 40+ hours every second week staring at a screen, doing bare metal wizardry on embedded MCUs[2] (and all the annoying admin stuff that comes with it), and that itch disappears...


[1]: Bit of Trivia --- the 75 Kombi that's currently my home has a Pi4 as its central hub, connected to 3 Picos to manage everything from controlling the lights, over monitoring the battery system and displaying the data on a bunch of hacked e-Paper thermometers, to automatically updating the map showing my family my latest travels. And then there are the Pi's I carry for work...
[2]: Love the task/event system of the Nordic nRF52 series!

Last edited by umisef on 01-Mar-2023 at 11:07 AM.

@kolla

Quote:

Amithlon/Umilator would have been awesome and survived a heck lot longer had it been open source from the start and not been dragged into the entanglement that is “commercial amiga industries”.

Counterpoint: Amithlon would never have existed if it had been open source from the start. I was, at the time, a "poor, starving PhD student" --- which is grossly inaccurate, but at the same time, is also to the point. Amithlon was not, and could never have been, a hobby project. The amount of time I spent on it would have been entirely unjustifiable for me without the promise of financial payoff.

In my case, the inaccuracy of the description meant that I had the luxury of being able to justify it with a promise of payoff, rather than actual payment. Yet when, due to circumstances best not dwelled on, the reality lagged the promise by a fair bit, actually having to put food on the table was one significant (albeit not the main) reason to move on, and spend time on things more lucrative (and more enjoyable).

There aren't many people who can (a) afford to work full-time plus on a hobby project and (b) are young/foolish enough that they think doing so is a good way to spend their days.

Back then, I was OK with (b), but failed on (a). 20+ years later, it's kinda the other way around --- before the pandemic hit, I actually had organised with work to take a year off in 2021. But I wouldn't have spent it bent over a keyboard; rather, I would have backpacked around the world.

Last edited by umisef on 01-Mar-2023 at 11:08 AM.

@all

I don't wanna be nitpicking, but do you (all) remember how the proposed AmigaDE was perceived, when first presented by Amiga Inc?

No matter how it would have been implemented: any vitual or emulated solution was unacceptable by people that (then) would have been far happier to spend thousands of euros on pointless, underpowered PPC machines running the same old stuff on a different processor.

By the way, a product like Amithlon/Umilator deserved to be commercially sold to user, since at its time, it was USEFUL: it solved the real problem to run at fast speed valuable software that ran on the 68K architecture, still using common hardware.

But time goes on and the once valuable software that ran on 68K has become just an historical curiosity today. You don't usually spend big money on mere curiosities. That's why I consider Amiga Forever a perfect fit, with a good price/performance ratio: it costs a few euros, but it pre-configures a perfectly emulated Amiga environment, with very simple management of third party software. A new game is released for Amiga? I load AF and it's running very well in no time.

My friends, considering as vultures everyone who produce anything that you *may* use is plain wrong. Not everyone is here just to give you new toys to play with, in their spare time. Development is a cost. The right to play easily is something is worth paying for. No wonder I've been happy to buy a TheA500 mini: it's a software + hardware solution after all, something I can touch and place on my desk.

It's not matter of selling things, but the price you pretend to apply to things. I would spend a few euros to download a Umilator ISO even today.

Last edited by paolone on 01-Mar-2023 at 10:51 AM.
Last edited by paolone on 01-Mar-2023 at 10:50 AM.

@Karlos

Quote:

Karlos wrote: I can't speak for Bernie but if I had to guess, Linux was chosen because it got the job done and allowed him to realise his ideas.

Precisely. Harald's original proposal actually was to build something like Amithlon on top of AROS. No, not using AROS inside the emulation, to replace AmigaOS, but underneath it, to do the job Linux ended up doing.

I don't know what the situation is these days, but back then, AROS simply wasn't up to the job. It was quite picky about what hardware it would run on bare-metal. It also lacked a lot of the fun MMU functionality I so enjoy abusing :)

Also, as Lasse put it,
Quote:

back then he was a linux guy

I started using Linux in '91, in the 0.95 days --- when "using Linux" meant something quite different from today. So ten years later, I felt quite at home inside the kernel.
Scarily enough, another 20+ years later, despite all my work having been done under linux for all that time, modern kernels intimidate me :-/

Last edited by umisef on 01-Mar-2023 at 11:02 AM.

@DiscreetFX

Quote:

Also it’s very obsolete, dead and the corpse kind of smells

That read a bit... harsh.

I'd rather think of it as being gracefully mummified; And mummies apparently have a cinnamon-like aroma.

Last edited by umisef on 01-Mar-2023 at 11:10 AM.

@umisef

Okay then, it’s a mummified dead rotting corpse that smells really really bad with hints of tobacco, chocolate and is worse than the bubonic plague.

Last edited by DiscreetFX on 01-Mar-2023 at 04:36 PM.

_________________
Sent from my Quantum Computer.

@umisef

Quote:

umisef wrote: @Karlos Quote: With respect, I disagree. I shall take that as a compliment.

You should. And I don't hand out platitudes willy-nilly :)

Quote:

However, given that I know what you are complimenting me on, and you don't, I equally respectfully disagree right back :)

As is your prerogative.

Quote:

This was never a well-designed and well-planned piece of code. Not a lot of planning is even possible, when most of the most interesting problems only become apparent after you have solved a bunch of other, quite complex problems.

I expect nothing less than reflection. If you were able to look back at code you wrote 20 years ago and say, "That's all awesome, there is simply nothing that could have been done better!" I'd be somewhat surprised as it suggests that despite having a formidable engineering skill at the time*, you somehow managed to not improve in the intervening years, despite the huge progress made in every aspect of hardware, software and engineering since.

* bear in mind that most otherwise perfectly capable software engineers simply wouldn't know where to begin with an endeavour like this. I count myself among those despite being relatively proficient.

Old software archaeology is not a zero-sum game. There is always value in the learning and even if 90% of the source were somehow irrelevant today (which I doubt since the intended guest hasn't moved on much even if the host has), lessons on how not to do something are as important as the "golden hint" on how to absolutely nail something.

Quote:

Amithlon has some of each --- some gems I'll happily reminisce about, and some mess that I'd probably despair over if looking at it again. But the thing is, even the elegant bits (or maybe even especially those bits) have very little to teach. The magic of those bits is in the journey, in all the wrong or clumsy approaches that preceded them, in understanding what those approaches lacked that the elegant one has, and maybe in developing an intuition for how to get from there to here. And none of that is in the source --- source only tells you what, it doesn't tell you the much more interesting why.[1] On a related note, sources don't tell you what could have been, but for the constraints of the time; And the flip side, what assumptions are needed for code to work, which may only be true in the code's specific context.

Understood. What I am really proposing, if I may be so bold, is that should you happen to reacquaint with the sources, consider making them available on GitHub or similar. As a software engineer, I have really bad old code that is cringe to look at but having lost most of it due to a slightly lackadaisical attitude to backing up (lost my local SVN before I could do that) I put everything I found on GH, warts and all.

Last edited by Karlos on 01-Mar-2023 at 05:22 PM.
Last edited by Karlos on 01-Mar-2023 at 05:19 PM.

_________________
Doing stupid things for fun...

@umisef

Quote:

umisef wrote: @DiscreetFX Quote: Also it’s very obsolete, dead and the corpse kind of smells That read a bit... harsh. I'd rather think of it as being gracefully mummified; And mummies apparently have a cinnamon-like aroma.

Zevulon the Great. He's teriyaki style!

_________________
Doing stupid things for fun...

@paolone

For the avoidance of doubt, the reference to Vulture was intended as a slight towards the various licensing trolls (for want of a better description) that would seek to capitalise instantly on something like a revised version, or shut it down.

I would gladly pay for a modern version where the costs were properly renumerated back to the engineer(s) responsible if that is the only way it could happen, even if I'd prefer it to go open source so that perhaps a community of like minded people could make it happen with someone like Bernie consulting.

_________________
Doing stupid things for fun...

@umisef

Quote:

umisef wrote: the 75 Kombi that's currently my home

Is that aussie slang for a T2 bus, or am I getting something wrong here?

_________________
- We don't need good ideas, we haven't run out on bad ones yet
- blame Canada