Quote:

One board...
This doesn't worry anyone else?

@Jim

What else did you expect ?

As long it is something easy to find and supporting a wide range of CPUs I don't see much of an issue.

More reason of concern would be how long it takes from starting on lets say this to release as that might put serious hurdles into sourcing such mobo (which will most likely be discontiued by that time).

On the laptop front things might get even harder, but besides from basing MorphOS_64 on a linux kernel (like AROS-hosted) I see really no alternativ with the limited resources.


But don't forget that is really just for the initial release (just like MorphOS at some time only supported one type of PPC Mac), and if one of the key developers decide that he really needs MorphOS4.2 running on his new 2025 MacbookPro I doubt the other would get their pitchforks to stop him.

Quote:

This is a rather old quote, and I think you're maybe misunderstanding me. I'm not saying that we'll only support one motherboard forever. I'm saying that we'll port MorphOS to and support running MorphOS on one specific motherboard (and one laptop) at a time. When that particular set of hardware becomes obsolete, or whenever we feel like it, we'll then pick a new set of hardware to support. Basically we want to avoid the current situation where we support such a large amount of different hardware that it becomes difficult to make sure everything is well supported.

Quote:

I'm surprised to read this kind of comment from you Jim. You have stated on many other occasions how much you trust the decisions of the MorphOS Dev. Team members, so what has changed in your point of view?

As you and I, and most other MorphOS users have stated in the past, the choices of which hardware to support by the Dev. Team members have so far seemed the best and most logical choices. I expect them to continue making wise choices for x64 hardware, just as they have always done with PPC hardware (with perhaps the only exceptions being support for the SAM460 and X5000, IMO ...).

Sorry guys, I tend to be somewhat neurotic about things that are important to me.

And I DO trust our developers to make the right decisions (they have aced it so far).

Although...if I had the cash I would buy at least a couple of Power8 boards for these guys.

>> the current situation where we support such a large amount of different hardware
>> that it becomes difficult to make sure everything is well supported

> the choices of which hardware to support by the Dev. Team members have so far
> seemed the best and most logical choices.

Seems at least some MorphOS team members are less confident about the team's past choices ;-)

Morphos on x86 will be not binary compatible like on powerpc because x86 cannot work in big endian mode.
It is stupid to repeat old mistakes.
Morphos on x86 should have not problems with memory protection, drivers, use of many core, porting.
It is time to cut off 90% of old API, leave only amiga graphics and gui and use unix as base.

Quote:

Yeah, like anything involving a group, there will always be some dissent.
And I myself still like the idea of sticking with Power.
But that might just be the "sore loser" in me.
I didn't really dig having to move from 68K to X86 years ago and still think PPCs are adequate for our uses.

Quote:

Are you out of your mind?
UNIX?
Never.
Should we ever go there (or Linux) I am out of here.
And we don't support native Amiga graphics right now.

We already have a good kernel.
And Ambient can be retained (and improved), along with some other elements.
Outside of that, we might as well let our gurus decide.

Quote:

I meant from the users point of view, not for the ease of supporting the OS by the Dev. Team. I am sure that many of the team members have mixed feelings, and regret supporting so many PPC models, as it makes more work for all of them. Their choices have made it easy for any MorphOS user to find affordable, (and in terms of NG Amiga choices, powerful) hardware all around them.

[ Edited by amigadave 06.05.2016 - 14:36 ]

I guess they figured that since most hardware are more or less the same supporting them would be pretty straight forward. But every hardware has it's own little quirks regardless. It's easy to look back in hindsight now when we know this.

Unless the developers have retained at least one of each platform they have ported to, I imagine it is difficult to assure that all previous systems work with each revision change.

Quote:

Quote:

Are you out of your mind?
You want to change API to not binary compatible and there still will be problems with software porting?
New system should be based on UNIX.

Quote:

MorphOS will *never* go POSIX/UNIX/whatever, anyone with knowledge of the MorphOS developers past view on this knows perfectly well that there isn't even a point in asking this question! It will never happen, and why should it? Sure, there will be some additions to the API, and some changes in rules and assumptions that affects applications environment, like no app will be allowed access to memory outside its allocated space, and that different parts of the system, and different parts of any threaded applications, can be executed on different CPU cores in an asynchronious manner unlike today. Etc, etc. But MorphOS was never about POSIX, never about making another Linux. My experience from past discussion from the MorphOS team, is that they would rather drop it and walk away than going in that direction. They are making *their own* operating system!

Quote:

The new system will be based upon MorphOS, I think this is a fair and realistic assumption.

Quote:

Knock yourself out. ;)


http://repo.arixfoundation.com/downloads/

Quote:

No,it should NOT.
That would result in a monolithic kernel, far lower startup
times, and and OS with no differentiation from all Linux distros.

If you want a UNIX base, use AROS hosted.
Don't request corruption of our OS.

Basing MorphOS_x64 on linux is the easiest and most obvious way
While we're at it replacing MUI with QT/GTK is the easiest and most obvious way
Compared to Wayland CGX is crap anyway

And now, the name "Morph_OS" just sounds stupid, lets replace it with something sensible.

I propose "Bluehat Linux" !!

While i also don't want it to use a *ix kernel, the idea is not too stupid again. I think Amithlon and the never released Umilator have shown a way to use a Linux kernel for the base job and still keep it Amiga. But I don't favor such a solution. I want the real thing.

But I would pretty much welcome some drivers optimized for a virtual machine like vmware, parallels or qemu.

Using Linux kernel doesn't have to make it another Linux distribution. Look at OS X: the fact it's partially based on BSD kernel doesn't make it a BSD clone.

The same would go for Linux. I don't understand these false assumptions...

Of course, the way that AROS chose of being hosted make it this way, but it doesn't have to be.

[ Edited by Leo 07.05.2016 - 22:40 ]

Quote:

The voice of "reason" finally! I agree that MorphOS could be based on parts of Linux, without making it just another Linux distro. With recent progress on AROS, or more specifically, AEROS, it runs fairly well and doesn't seem like a "hosted OS". Maybe we will be able to claim that MorphOS for x64 is Amiga and Linux done right!

Wouldn't it be a great thing for the MorphOS Dev. Team, and all MorphOS users, if MorphOS for x64 could take advantage of most, or all of the Linux drivers for x64 hardware, AND be able to run Linux programs (perhaps with a wrapper, or some other tweak)? I say yes! YES!

If the MorphOS Dev. Team went this direction, it would save them months, if not years of coding work, and I am confident they can pull it off in a manner that acts like our existing PPC MorphOS3.9, by hiding all user action parts of Linux they use, and replace it with the MorphOS GUI and organization we are all used to and appreciate from our background with Amiga computers. Even if the MorphOS Dev. Team completely changes lots of things that Linux distro's currently use, or have set up a certain way, MAYBE they can still take advantage of Linux drivers, and development tools (perhaps with modifications), and make it easy to run Linux software, or make it easier to port Linux software to MorphOS for x64.

It will be slower, but someone already pointed out that MorphOS for x64 will be slower no matter what they do, because they will be adding memory protection and hopefully SMP. I believe that the speed of MorphOS for x64 will still seem faster, due to the faster hardware we will be running x64 MorphOS on. An i7 CPU is how many times faster than a 2.7GHz G5? 10x, 20x 50x???

The MorphOS Dev. Team can work with the AROS developers and take what ever is worthwhile, (if anything) and return perhaps some improved parts back to the AROS programmers and users, if that is possible.

To my mind, this is the only logical way forward, which would help us stay close to the leading edge of other computer platforms, and also allow faster updates to new hardware, when the initial x64 hardware choices are no longer available. If the MorphOS Dev. Team write all of their own drivers from scratch every time they need to update to newer hardware, it will limit us to fewer choices, and take tons more time, every time the existing hardware becomes unavailable (and in the land of Intel PC's, that seems to happen really fast, like 18 to 24 months).

This route may also entice some Linux programmers to join our ranks, since the development tools should be familiar, if not exactly the same tools.

[ Edited by amigadave 07.05.2016 - 16:29 ]

I'm not saying they should go for Linux (or any other open source modern kernel), what I'm saying is that going Linux as the kernel doesn't mean you cannot write an OS that doesn't feel like an Amiga and retains Amiga-like features (MUI, Datatypes, Volumes,...).

That said, I'd be surprised if they went that way.

Thought I would let this get bounced around a bit, as my inital response was tainted by my emotional reaction to this idea.

OK, so why not?
Because, unlike UNIX or Linux, our OS currently uses (quite sucessfully) a microkernel.
This feature is relatively uncommon as it is most commonly found in OS' used in process control applications.
It's compact and efficient as opposed to the Linux kernel (which grows in size with each revision).
And microkernel based OS' are inherently more stable as features that in Linux are incorated into the kernel are often placed outside the kernel in a microkernel OS (where they are less likely to crash the system).

In short, the kernel we currently use is OURS. It was created by core developer Ralph Schmidt. Its basic design is quite different than a Linux kernel
It is tight, clean and rather elegant.

I do NOT favor throwing away core features, that offer a benefit and differentiate us from the masses.

I could be biased, as the last operating systems I used on the 68K were microkernel based.
But, there was a big advantage to that in the '80s and '90s, and those bonuses are still there.
You will NOT hear me join a chorus of voices that would suggest we resort to such a cobbled together hybrid.


[ Edited by Jim 08.05.2016 - 11:08 ]

Quote:

Argh!
Why not just move to Linux?
You could just clothe it in a MorphOS theme and be done with it.

Why not base NG MorphOS on TempleOS instead?

AFAIK all the things needed for a modern kernel is already there in Quark. Sure, you get a lot more drivers in Linux, but will it still be it's own thing? I don't think so. I think it's the same reason why Haiku voted against going for Linux too. It would be less authentic and less original.