Quote:

Good point. I'd support that with my money.

Any team members reading this that would prefer this option?

Quote:

+1. I don't need yet another ppc hardware. We have loads of Macs to choose from and Sam460 and soon X5000 so there is enough hardware base. Such a bounty for another board (with no promise to be ever supported) is a waste of money.

I would put some money on x86 or x64 port though.

> the pre-release version of the Pegasos mainboard was primarily a tool for testing MorphOS.

But this kind of opposes the idea that the Betatester was "combinations of pre-release versions of the Pegasos as well as pre-release versions of MorphOS", doesn't it? Either the tester is the Pegasos board that is used to test the testee MorphOS OR the tester is the combination of Pegasos board and MorphOS (which begs the question what the testee is) ;-)

> there was no general expectation that testers would have the required capabilities
> to correctly identify and report hardware bugs

Testers would report bugs (e.g. data corruption) by describing the observable end result and what they did to get this result, so that it's repeatable for the receiver of the report. They wouldn't have to know whether it's a software bug or a hardware bug in order to file that report.

I wonder if the MorphOS for x86 bounty idea should have its own thread.

[ Edited by Everblue 19.11.2015 - 14:21 ]

Quote:

Better wait for any developer with any statement on that, otherwise we will end up with another useless topic full of speculations and we probably had many of them already regarding MorphOS na x86 ;)

Quote:

Agreed.

For me it sounds like there is some goodwill regarding this project both by Freescale (providing the SoCs at a realistic price) and Genesi (creating the design "at cost").

This could lead to a board with performance close to the X5000 while costing less than a SAM.
I see a whole lot of "point" in that.

Quote:

Why not? I like competition. It has been parroted that it is impossible in 2015 to make a PPC board and sell it for a reasonable price. Even here on Morphzone we are being told A-Eonkit are borderline martyrs with their current pricing. Let's see some competition back in the market to find out if that is true.

BTW the Tabor will be a trainwreck IMO. I can't imagine any scenario other than A-Eonkit losing gobs of money on R&D with that board. I'm rather surprised you see any value in it at all. I guess at this point your stances are just wanting to support the people behind A-Eonkit, and not the actual product.

Quote:

Supposedly there is a clamoring by MorphOS users for "new" motherboards. I don't see it, but certainly a reasonably priced and reasonably powered motherboard would be preferable over what is currently being proposed.

Quote:
Not at all. Acknowledging that a piece of hardware required for testing software is not ready to be sold to end consumers (because of known bugs in the included Northbridge chip, for instance) does not automatically imply that there is any expectation that having it evaluated and tested by external parties would yield any new insights of notable value.

Quote:
The tester was the hardware, which, as previously explained, was a complete system initially.

Quote:
In order to be any real help, they would have had to know that and report to the appropriate parties in a very timely manner. It can take many months to figure out that cases of data corruption were not caused by software, especially if it is under heavy development, but by hardware instead.

You make the proper reporting of bugs and confirmation of issues sound completely trivial, which they are not.

Quote:

Where did this come from?

Quote:

And is it possible to keep it "reasonably priced and reasonably powered" in such a small quantities? I doubt it.

If you guys were already running MorphOS on your preferred architecture, then what? What would you be able to do that you can't already do?

Quote:

If we would get more modern browser (and I believe we would) then I can easily browse, without worrying about lack of memory or that some crap page can ruin my whole work with browser crash or whole system crash.

Quote:

What kind of quantities did the Pegasos I and Pegasos II sell? I got a Pegasos I w/April 2 ~12 years ago for $299 with teeshirt, MorphOS 1.3x and several games. IIR the Pegasos II went for ~$299-$499 depending on the config.

For something more recent, I will (hopefully *) be getting a C64 Reloaded for $140.00 soon.

https://icomp.de/shop-icomp/en/shop/product/c64-reloaded.html

While I understand this is lower tech running at a lower frequency, it is even lower volume by a far amount, and hand assembled with rare NOS chips.

* Hopefully because they are hand assembled and only sold a few at a time. They are in such high demand, relatively, that Jens could jack the price up exponentially, but he doesn't. I say *hope* to get one because as soon as they are listed for sale they are snapped up. I'll have to be nimble.

Quote:

Primarily I need more cpu power and RAM for convenient browsing, hd video and demanding pdfs. Especially the RAM issue is important. It cannot improved (much) w/o breaking up binary compability. AmigaOS is virtually bound to 31 bit addressing. Going to a 64 bit (even going to use the full 32 bit or 36 bit many ppcs support) address space is a major change which will lead to a binary compability breakage.
And while breaking binary compability add SMP and full resourse tracking and (maybe) MP, too.
In short: make a MorphOS NG.
And I don't care which ISA is used as long as it warrants fast, affordable and easily obtainable hardware. And in my book this is best fullfilled by x64 hardware.
Additionally ISA related/dependent developments (like JS JIT) are easier on a widely spread architecture.

On the other hand: MorphOS currently runs on ppc. A new modestly powerful board with a remotely reasonable price has its charme. I would actually welcome that. But only if it doesn't hinder MorphOS NG development.

>>> the pre-release version of the Pegasos mainboard was primarily a tool for testing MorphOS.

>> But this kind of opposes the idea that the Betatester was "combinations of pre-release
>> versions of the Pegasos as well as pre-release versions of MorphOS", doesn't it?

> Not at all. Acknowledging that a piece of hardware required for testing software is not ready to
> be sold to end consumers [...] does not automatically imply that there is any expectation that
> having it evaluated and tested by external parties would yield any new insights of notable value.

This may be, but I don't see the relation to my question, which essentially was whether what was called "Betatester" was the pre-release Pegasos board (as insinuated in some postings) or the pre-release Pegasos board plus pre-release MorphOS (as insinuated in other postings).

>> Either the tester is the Pegasos board that is used to test the testee MorphOS OR the tester
>> is the combination of Pegasos board and MorphOS (which begs the question what the
>> testee is) ;-)

> The tester was the hardware, which, as previously explained, was a complete system initially.

"Hardware" implies the Betatester was the board, whereas "complete system" implies the Betatester was the combination of hardware and the OS.

Quote:

Reasonably powerful?
Yes, I think that can be done.

Pricing?
That is a trickier subject.

For a Mini-ITX board featuring a 1.4 GHz T1042, just slightly higher than the Peg2 price quoted seems possible.

But my own preference would be something more like a 1.8 GHz T2080, and that CPU costs what about three of the T10XXs goes for, and the board would likely be more complex as the T2080 could support more expansion slots.

I'd pay about $350 for the former and $700 for the latter (possibly higher for both), as each (if SMP enabled) would outperform the X5000 (with the second board having a huge performance advantage).

The first build is the easier to prototype as Freescale has similar reference boards.
The second spec is unlike any existing boards I know of.

Frankly, I really don't think CPU power is as much an issue here as is SMP and addressing a larger memory range.

And if MorphOS is to move to X64, this could provide a more gradual staging for that move.

Both of these ideas offer significant advantages over the products we are likely to see from A-eon for the next several years.

So, the question really becomes, if A-eon thinks its going to sell at least 1000 Tabor boards and at least 500 X5000 boards, would there be a place for a lower volume board designed to cater more to OUR market than these?

I am pretty sure a superior design is possible, AND that there would be enough buyers to justify runs of, say, 50 boards at a time.

But it should be understood that these ideas are not born from practical concerns, but actually address what I and a few others would like to have available.

And since I have a history of helping make what I want for myself available to others, I think 'why not do it again?'.

Quote:

OTOH none of those you mentioned makes a lot of sense, the Sam is both underpowered and overpriced, the X5000 is not state of the art either, even for PPC, not when it comes to performance, it does have its fair share of "stupid" attached to it (the "X" stuff that not even A-Eon themselves knows the purpose of) and it won't be cheap, and the Tabor is kind of broken by concept, no matter its price. And none of the X5000 or the Tabor is actually available.

A powerful, well designed, reasonable priced motherboard could perhaps turn the table, making *the others* redundant.


Quote:

I participated in the last attempt of a bounty-funded motherboard development. It didn't work.


Quote:

Personally, I will support *products* with my money. Or at least clear promises. Or at least tangible projects.


Quote:

I agree, at the very least there would have to be a public commitment. And there are development boards one could use to evaluate or get development going before a motherboard is ready.


Quote:

I see no need for a x86 bounty or a thread about a x86 bounty. It looks like it is happening anyway.


Quote:

Could very well be, but information is scarce, only one or a few vaguely written forum posts from BBRV. Could mean anything, more probable is it means nothing.

> each (if SMP enabled) would outperform the X5000

...with applications that can take advantage of more than 2 threads. And even then the T1042 would outperform the P5020 in the X5000/20, but not the P5040 in the X5000/40.

> The second spec is unlike any existing boards I know of.

There is the T2080QDS board but I haven't been able to find out its form factor.

> Both of these ideas offer significant advantages over the products we
> are likely to see from A-eon for the next several years.

You're ruling out that A-Eon will commission development of a T1-based or T2-based board after Tabor?

@TMHG

The _only_ real reason to buy any of these boards is to run AmigaOS 4 or MorphOS on them. It's not about power or anything, only that you absolutely want to buy new hardware from a tiny vendor instead of buying more powerful and much cheaper used PPC Macs. In AmigaOS 4's case the reason is obvious: no used PPC Macs. For MorphOS it makes a lot less sense.

There are a few who will buy these boards for the fun of it, but they will be few. And they will be available long before this new hypothetical board will leave the planning stage. By this time, those who really want a PPC board will own one. And 2/3 of these currently selling ones will already be supported by MorphOS.

People complain all the time that PPC Macs will eventually break down, which is of course true, but they at the same time buy much older Amigas (who are more prone of breaking down) without making a fuss about that. And people make no fuss about the fact that these new boards are made by a tiny company with little money to test and develop, which probably will make these boards more prone to having problems than boards made by huge corporations. People don't think about that either. There is a chance that my Mac Mini might outlive the X5000 for this reason.

> 2/3 of these currently selling ones

There's only 1 of those 3 been released so far. And even that 1 seems hard to come by judging by the vendor's availability status.

> these new boards are made by a tiny company with little money to test and develop, which
> probably will make these boards more prone to having problems than boards made by huge
> corporations. [...] There is a chance that my Mac Mini might outlive the X5000 for this reason.

The Nemo board has been in the hands of end users for almost 4 years and I've yet to hear of any problems in terms of design faults or bad build quality. Varisys seems trustable here.

Quote:

Luckily we have a lot of people willing to support this kind of projects.

Quote:

G5 port would probably happen as well but we made it to appear quicker. That's great example of interaction between bounty, community and developers.

Quote:

Hence the:

Quote:

Acube has stated that they will keep making SAM 460's.

Quote:

Except for Epsilons broken X1000 you are right. I'm not talking about bad design here, only that bugs do creep in and mistakes are made in the building process. But the testers of the various PPC boards aren't that many which makes it harder to spot them. Apple is a huge company that lives on the image of "things just working" and "rigorus quality control". A-Eon and Acube caters to a hobbyist market of a couple of hundred users. I'm not saying at all that the two latter is not doing a good job with what they do, just that they lack the resourses Apple had for making good hardware.

And 4 years isn't a long time. 10 years not so long either. Which is the age of my Mac Mini.

[ Edited by Yasu 20.11.2015 - 16:19 ]

Quote:
This is a distinction without a difference. In order to be able to test software, that software would obviously need to be provided to people who are expected to test it - along with the required tester hardware, which is exactly what happened.

Of course, once received, the tester hardware would stay the same during the testing phase while the software part would continously receive updates and change. Therefore, choosing a name that focuses on the single fixed element of a total package and its intented function would by no means be an unusual choice.