Quote:

I was thinking a MorphOS port to ARM *could* have some point, because ARM runs in beg endian mode, too. This would help to integrate a 68k emulation just like under ppc MorphOS. But as it seems the big endian mode of the ARM is not perfect (but I haven't read too much about that). The "native mode" of ARM is little endian.
And if ARM doesn't offer the full benefit of the big endianess, then I do not see much of a point in porting MorphOS to ARM, *then* a port to x86 would make pretty much more sense (x86 will continue to be the main cpu architecture for teh foreseeable time).
Well, next is the Mac mini port. We'll see what comes afterwards.

Addition to my earlier post why a MrphOS port to ARM *could* still make some sense.
As it seems ARM is entereing the netbook market this year. For ARM there is virtually only Linux available. And as we all more or less know Linux suxx (well, some ppl may disagree, but we all may agree that many ppl think that Linux suxx some serious donkey balls). Those ppl may be in some demand of another OS. Windows will proabalbly enter the ARM netbook market, too (once there is a serious market). but w/o the advantage to have all those x86 apps around. So, the biggest benfit of Windows will be void there.
Thus, on this "new" architecture there might be a relatively high share of ppl who may be open for an alternative OS. Question would be: Would they get satisfied by a MorphOS ARM? I am undecided. I think MorphOS is better than many ppl say, but it surely has its weaknesses.
If I would be a gambler I'd probably propagate a fork to ARM. But lucky me, I am not the one to decide .

@ Zylesea

Quote:

More importantly, you are not one of the developers who would have to develop and maintain all these forks

On a more serious note, it does not take a psychic to predict that most desktop computers will be sold with more than 4GB of system memory in the not-too-distant future. If I had my own OS and anticipated a labor-intense and time-consuming port to a different processor architecture, 32BIT ARM chips would seem like a rather backwards choice to me.

Quote:
Well, the Amiga has been moving backward since a decade anyway... The contrary would be surprising...

The OS is also moving backward, with compatibility (ie: past) in mind, not with the future in mind. In that case 32BIT ARM seems like a perfectly good choice... And Amiga applications don't need that much RAM anyway, right ?

> as it seems the big endian mode of the ARM is not perfect (but I
> haven't read too much about that).

Search for "true big endian" (as opposed to just "big endian") and "true little endian" (as opposed to just "little endian").

@ Andreas Wolf

Yes, I think i got it now. ARM chips are able to load and process big endian data, while addressing und such stuff are always little endian. I gues it is pretty useless for MorphOS (in the regard to yield a transparent 68k emu as with ppc).
We're on ppc, we're doomed. Well, doomed, but havig fun, I'd say

Quote:

That's the spirit!

Quote:

IMHO x86 makes sense in a desktop context, but MorphOS can't compete on the desktop market; the OS lacks many important features, and on top of that it lacks 100% of the applications.

MorphOS is a low footprint, extremely lean OS that can really do magic to the user experience on cheap, low performance hardware. This is where MorphOS can shine, this is where it can make sense. This is the strengths one could capitalize on!

One of the things I like with PPC architecture is that it scales from very low performance SoC designs up to quite decent performance "desktop" CPU's. The same MorphOS can happily run on a $15 "G2" CPU @ 400MHz as well as a Pegasos2 or Mac G4 @ 1.5GHz. We "Amiga Nerds" likes the *power*, since we are using MorphOS as a desktop OS, while the low power chips (like the 5121e) is what could make sense for MorphOS in a commercial context. The PPC offers it all.

However, it looks like the future of PPC isn't too bright anymore, so eventually MorphOS will have to jump somewhere else. Unfortunately, the ARM doesn't scale as high, and the x86 doesn't scale as low. Commercially, I can see how MorphOS potentially could make sense on the ARM architecture, but ARM driven netbooks and other kind of devices may not be what a MorphOS *desktop* user is dreaming about, because the lack of raw power? The x86 offers the power but it doesn't fit in the ARM marketplace, and commercially I can't see how MorphOS could possibly compete on the desktop market.

So if MorphOS would jump to a different architecture, and if the ambitions are above forever being merely a hobby project, then I would vote for ARM.

Well, what if a MorphOS ARM fork would simply cut the ties to the 68k legacy altogether and start with a clean slate? Could you really reach a good enough emulation speed on a chip with such low performance anyway?

BTW (speaking of restarting with clean slates), Anubis is going ARM:

http://bbrv.blogspot.com/2009/03/anubis-on-aura-on-arm.html

> Anubis is going ARM

Yes, already known for 1.5 months:

"That same package is ready for the i.MX515 [...] The Anubis Dev team also wants to develop their system for this platform."
http://bbrv.blogspot.com/2009/02/amigaos-morphs-to-pegasos.html

"Michal Schultz [sic!] has proposed to port [...] Anubus [sic!] to our new i.MX515 based hardware"
http://bbrv.blogspot.com/2009/02/communities-build-volume.html

http://projects.powerdeveloper.org/project/imx515/736

:-)

@ takemehomegrandma

Quote:

You have been repeating this for quite some time. However, running a successful business is in most cases not at all about "beating" your competition but rather finding your niche.


Quote:

In simple terms, developing a piece of software as a business requires you to generate sufficient revenue to cover all costs and preferably still have enough left to build up cash reserves for weak periods or emergency expenses.

That is all you have to do.

Let us assume that you focus your expenses on what is really important. No private jets. No first-class flights from the US to Europe and back. No fancy sports cars. No expensive offices. No redundant lawsuits.

How many sales do you need in order to turn your hobby into a profession? 100 million units? 1 million units? 100.000 units?

You can do with a lot less, actually. For the sake of this discussion, let us assume your software product is an operating system that includes a suite of every-day applications and runs on fairly common hardware. If you manage to sell a mere 2.000 units per year at $100 each, you generate 200.000 USD in revenue. At an admittedly modest average income of 3.000 USD per month, this is sufficient to pay 5 full-time developers or 10 part-time developers at half the salary or a combination of these. At the end, you would still have $20.000 left to pay other expenses or to save up.

Now, there is no need to pick apart small details of my example which is obviously very simple and abstract .The point I am trying to make is that you can in fact run a sustainable business with comparably small sales volumes as long as you can keep them at a constant level thanks to continuous product innovation.


Quote:

What we do know for a fact right now is that there are at the very least hundreds of people in the world who are willing to spend half a thousand dollars and more in order to buy custom hardware and a keyfile for the single purpose of being able to run MorphOS and compatible applications.

Would it be possible to raise sales numbers if the supported desktop hardware was more widely available, possibly cheaper and overall had a better price-performance ratio? Well, nobody can say for sure.

However, this - purely hypothetical - scenario sure sounds considerably more realistic than any vague dreams about large OEM deals for low-performance hardware (with cut-throat pricing, thus tiny margins for system software), embedded usage, and so on.

While having grand visions can be sometimes fruitful, ignoring your existing custumer base is almost always stupid.

Quote:

Hopefully the Mac mini will soon provide us with a case study whether a more widely avaiable hardware with better price/performance ratio will yield to more sales or not.

Not bene: Yes, the Mac mini is also quite niche and not new, but at least I guess we (or to be more precise: the MorphOS-Team) will get an idea how market demand for MorphOS for easily availabile and not overpriced hardwrare will be.

Quote:

Indeed. But how would you define a plausible niche for MorphOS on the commercial x86 desktop market?

takemehomegrandma:
Which Netbooks are most likely to survive
and thrive? Standardized x86-based or
customized ARM-based?

I know what I think.. At least the x86-based
can run Windows and Ubuntu, which makes
it even more appealing to many users.

Quote:

You mean "standardized" x86-based or "standardized" ARM-based (and I don't see any real standards as of yet BTW)? And I am not *only* talking about netbooks BTW, ARM is being used in many other types of devices as well where MorphOS could make a difference. Much of the ARM market is far out of reach for the x86, but maybe that will change further down the road. I agree about the Windows part though; the Windows trade mark and familiar GUI alone is a great competitive advantage...

Takemehomegrandma:
The Atom-devices are based on a Netbook-standard
set by Intel. Via is trying to get their out.

You could use any of the millions of x86 Netbooks
produced each month.

Could you please give a few examples of other uses
than Netbook where MorphOS and ARM would fit?

No, the ARM is not enough for high-def media players..

Quote:

The i.MX515 ARM-devices are based on a "Netbook-standard" (let's call it "reference design" instead, for that is what it is) already set by Freescale.

Millions of ARM netbooks will be produced, I see no reason to doubt that. The 2009 holiday shopping season is the target, and I guess we will see after that, like towards spring 2010. AFAIK, the Atom netbooks doesn't compare in cost/profit margins, power consumption, etc (but obviously has the power of Windows). And one other thing, the "netbook boom" has yet to come (if it ever does, many predicted the boom to take place a year ago already, but the *boom* hasn't really happened yet).

Other devices would be LCD/Plasma TV's (720p is, and will be, the DTV HDTV standard of importance), PVR's, STB's, handheld devices (PDA's, phones, etc), In Shop Displays, Info Kiosks, On train/in flight entertainment, or a fuckload of other consumer electronics devices where ARM is used. I mean, give me a TV with a HDD, USB, Ethernet (and optionally WiFi and Bluetooth) and *MorphOS* and I would die with excitement. Think of the possibilities compared to the locked down to a bare minimum "OS's" in these devices now. And I look at all these kind of devices and I ask - Where is the x86?

I actually agree with Matt above, and further suggest that there is a strong motivation for a migration to ARM, as PowerPC is fading into a dead end and ARM is ascendent.

The irony of course being that this is the first time I'd turned on my G3 in months to say this. And of course Neko is going to turn around, insult me, and we get back to our usual routine.

@ironfist

Actually, you're wrong on a few levels.

1) VIA is in the process of releasing their second-generation netbook reference. VIA's cloudbook reference is one of the oldest netbook platforms out there, pre-dating the Atom based designs.

2) ARMs are used in high-def media setups already, so you are misplaced there. Please refer to the Marvell and TI offerings, and the upcoming nVidia designs, off the top of my head. I believe others are working on it as well.

3) Commodity development boards are already available. I've purchased a Beagleboard myself for some IO experimentation. When you can get a 1Ghz ARM platform for $100, the times, they are a changing.

|||

[ Edited by Velcro_SP 06.07.2011 - 19:53 ]

Quote:
I don't think anyone is suggesting being tied to a Genesi product, for-se, but that migration from the PowerPC to ARM platform would bring far more benefits than drawbacks. Genesi does not come close to offering even a fraction of the ARM systems on the market. You have development boards from Marvell, Samsung, Freescale, TI, all available to work from. Look at the Beagleboard sometime, an embedded joy for $150.

My concern with PPC remains the lack of vendors for key components, and the reliability of them. While the CherryPal looks good, agreed (and I might buy one myself anyways) what happens beyond that? The PPC based Macs are aging, the PPC amigas even moreso. There is not a huge development pool of vendors to supply us with the necessary hardware, and those that do often times lack the leveredge to make full use of what they have. ARM does not have any of these problems, and the only real problem comes one of perception and upper scalability, only coming in speeds of up to 2Ghz. But, considering we're limited to G3 and G4 era PPC's anyways, there is no loss there.

> ARM does not have any of these problems, and the only real problem
> comes one of perception and upper scalability, only coming in
> speeds of up to 2Ghz.

ARM up to 2 GHz? Where? I see Marvell currently at 1.2 GHz and Qualcomm (not released yet) at 1.5 GHz. (*)

> considering we're limited to G3 and G4 era PPC's anyways, there is
> no loss there.

G4 currently scales up to 1.7 GHz (MPC7448).


Edit:
(*) Also Intel are currently at 1.2 GHz regarding their ARM processors.

[ Edited by Andreas_Wolf on 2009/3/23 4:10 ]

Quote:
Just looking at the roadmap. The fastest ones I've seen in person have been Marvell's 1.2Ghz dual core model (78200). But, consider for a moment, the prototyping kit for the 1Ghz Marvell is $99, and think of the options from there.

2Ghz ARMs will be shipping in the next year and a half, while G4's have stalled, same speed now as they were years ago. I just am tired of the dead ends, the bickering, the halted starts. When we are looking at the 7448, a chip now 4 years old, and that is the best we can come up with, we are in trouble.

Quote:

I understand that MOS should not be tied to a dead end CPU and please don't take what I am about to say in the wrong way or context, but one of the things I would hope the MorphOS community would try to avoid is the short life cycle that is perpetuated in the Windows world where you are expected to upgrade your computer every 1-1/2 to 2 years, or less, just because there is a faster CPU available, not because you need it to do what you normally use your computer for. With the increased efficiency of MorphOS we will always be able to do more with less computer resources, just as the Amiga did for years, and as such I would hope that our life cycle would be greater. Maybe 3 to 5 years between having to upgrade machines.

Just a rambling thought, for what it is worth.

[ Edited by amigadave on 2009/3/19 18:08 ]