@ CISC

I think that you, sooner or later, *will have to* choose a future architecture path:

1. ARM (Mobility, consumer electronics, media devices, etc)
2. x86 (Power, stationary desktop computers, servers, traditional laptops)

And if you agree that it's plausible that you indeed will have to make such a choice at *some* point down the road, then don't you agree that it would make more sense to make the decision *sooner* rather than later?

Of course, a third option of choice also exist (and that is the option to not make a choice at all):

3. PPC only, always and forever.

However, that doesn't seem like a path towards any kind of future, does it? Or if you think it does, could you please explain how?

> Click Products->32-bit Microcontrollers and then from the menu on
> the left of the new page, click 32-bit Automotive, and see i.MX..
> every i.MX family.

Thanks, I see now. Didn't notice before. I guess you know what list I mean where only some i.MX processors are targetted at automotive use. Seems to be the kind of usual information inconsistency Freescale are noted for ;-)

> I'm not going to argue with you, you have too strict of a worldview
> to take in any information. Freescale do not care what you use the
> processor for, as long as you can use it.

I think you misunderstood my intention. I did *not* talk about what *I* think a certain chip can be used for if one wants to use it that way, but about what use *Freescale* themselves explicitly target that chip at. That's quite a difference, and that's why I countered your claim that *Freescale* would be targetting the i.MX515 at automotive use. Whether it's the case or not is unsure, as there are obviously contradictory information on the Freescale website. Fact however remains that the i.MX515 product page only lists netbooks and nettops as target applications.
And just to illustrate that you misjudged my intention let me tell you that I would love to see a certain desktop operating system on a certain embedded image processor. Not quite fitting, right? ;-)

> The MPC5xxx exists right now I am sure only to please auto
> manufacturers who have an already-existing very large codebase.

Which specific actual i.MX processors do you think could substitute the MPC55xx/56xx?

Quote:

Or

3 Power familiy (from mobile phones to Super Computers)

Quote:

There are a couple things automotive companies want; CAN, MOST and FlexRay.

The i.MX35 implements the CAN bus and it's a very reasonable little chip with equivalent performance to the MPC5121e (just without the dumb multi-port memory controller).

MOST is a horrible automotive media communications bus which makes absolutely no sense in the modern world, but it has to be around for compatibility.

FlexRay is pretty new and replaces CAN with a higher performance bus, and most automotive products with display will end up using something different.

If you're designing a car from scratch and not just plagiarising chassis from old models and components from old models, any i.MX chip would do the job. For components such as OnStar (GM), MPC5xxx is already well in use and needs to be supported. For vehicle safety, you cannot just swap chips inbetween models..

@CISC

MorphOS does not need to retain compatibility. Do you want to know what Freescale did when we told them we had an OS that had m68k compatibility on the Pegasos?

They laughed. It's an absolutely pointless feature. Apple shipped MacOS X with a "MacOS 9 in a window" emulation rather than drag m68k emulation to X. They shipped OS X for Intel with Rosetta.. but that is far, far gone now (check out the stuff for Snow Leopard, they say it's 7GB smaller - that's 7GB of PowerPC fat binaries and compatibility layers and emulation cut down to only support 64-bit Intel systems)

Whatever chip MorphOS runs on, it should be one where MorphOS can become more than an Amiga Workbench clone that runs applications which stopped being compiled 10 years ago. I can't think of a single component - besides translator.library/narrator.device and ARexx, the latter of which I recall is your project - which may still require something being emulated on the m68k. The whole m68k emulation thing also adds a tremendous overhead on the API.

I've been involved with a couple Amiga-clone OS projects over the years and by far the best one was Carsten Schlote's CaOS project at Met@box. The thing was damn sweet; take AmigaOS and take out all those functions that did things the old way, and just do them the new way. No SetPatch+NewSetPatch+NewNewSetPatch+whatever, just one function that does it. You clean up the API and everything works so much better.

Ditch all the Amiga Blitter assumptions and interleaved bitmap support out of the Graphics API. Throw away the 50Hz power supply clock! Actually implement real multi-user support and memory protection.. rework the input APIs and support Unicode across the entire system without weird hacks. Bundle a web browser class as standard. Stop being an emulation "box" and start being an operating system.

A lot of those concepts got dragged into MorphOS by Stefan Stuntz for MUI 4 (which got rid of a bunch of libraries which caused some compatibility problems - amazing how many people used the "private" muigfx.library :)

Maintaining compatibility is bloat, and if you market MorphOS outside of the <500 people who really want an Amiga clone through rabid zealotry, it is not required. MorphOS has an opportunity to become a top quality embedded operating system with a significant market share, but it will never be this way if you follow the Amiga development model and maintain binary compatibility.

If MorphOS doesn't want to go in this direction, then I guess you will eventually stop getting registrations and bounties and everyone that matters will move to AROS, because for a community-written open source and "libre" project, they have a much better business model.

Nonetheless, I wish you luck with your diminishing market.

@Neko

Ouch...such a bitter post...

Anyway, I wish you good luck with your new Efika. It would be interesting to know where most of your sales from the original Efika came from. Because I know many people bought one just for the fact that MorphOS ran on it and no other reason.

Maybe you had way more "embedded" and "industrial" orders for it, making any MorphOS sales like a drop in the bucket, but maybe not.





[ Edited by HammerD on 2009/8/27 14:14 ]

@HammerD

It's not bitter at all; it's pragmatic and realistic. They say the truth hurts, but most of the time it is just that - hard to swallow for people who "believe".

Sales for the Efika specifically for MorphOS were very low. A lot of this is due to the delay of nearly 18 months to 2 years before a stable version of MorphOS was available for Efika.

If we take the maximum serial number reported here in July (533) and assume a good quantity bought it for Pegasos, or Pegasos+Efika (=2 keys) or Pegasos+Efika+A4000 (=3 keys) in order of probability, then you're barely making up 2% of sales. At the top end, 5%. I just checked my tables and it is not statistically significant any way I work the numbers.

In the end the Efika 5200B was not produced in large enough quantity to be a viable market for any OS - it would certainly net less than 1 Million EUR through registrations if EVERYONE bought a MorphOS license key, even if they all bought it after the price went to 150 EUR. Actual sales are less than 1/10th of that value.

(remember this is more expensive than MacOS X or Windows 7 Home Premium)

1m EUR is tantamount to paying 10 engineers for a year, after that, you would run out of money. If you take the 1/10th value above as a conservative estimate of how much money can be brought it for MorphOS, it basically means after tax and expenses, Ralph gets to eat ramen noodles, and everyone else continues to work as a hobby project with no remuneration.

This is not a viable situation for a company to be in that creates an operating system.. the Efika MX cannot afford to wait 2 years for an OS port. What it could do is support an embedded application for MorphOS which is sold on the side for the Amiga Desktop Market.

If you want me to be even more realistic: why would anyone buy MorphOS if AROS was free?

[ Edited by Neko on 2009/8/27 19:51 ]

@Neko

Most customers perceive free as a cheap substandard product with limited or no support. Example, Red Hat Linux does very well as a paid OS but free Desktop Linux struggles to gain any Windows market share. Apple repackages FreeBSD with NeXt wrapper and sells millions on thier hardware. No such luck for FreeBSD. Mac OS X is nice but so is FreeBSD which it is based on.

@discreetfx

That couldn't be any further from the truth!

That fact is that a lot of people see the free operating systems not as cheep, but as operating systems that have a GREAT deal of support.

Look at linux in general, how many people are supporting linux compared to Windows? I really don't know the answer to that, but I am sure it is a lot more!

I seriously doubt that ANY operating system will ever gain a decent share in our windows market.

Mac OS X is based on FreeBSD? I bet Steve Jobs would have something to say about that. Mac OS X is based on the NeXt operating system. NeXt was more advanced than BSD back in the 90s, and still is!

I make maps for a living and the software I use runs on Windows....I have no choice.

As far as fun stuff, I prefer the Amiga platform, I always have.

I can't believe that any hardware manufacturer could rely on the Amiga market.

Morphos is too expensive. I have no intention of paying 150 eur for an os that doesn't even have half of the features of an OS the same price(Windows).

I do love AmigaOs and MorphOS, I just think things have taken a wrong turn......

This is a sad discussion.

J

@jclausen


http://developer.apple.com/opensource/


It's based on NeXt & FreeBSD.


From

http://developer.apple.com/Darwin/


"Beneath the appealing, easy-to-use interface of Mac OS X is a rock-solid foundation that is engineered for stability, reliability, and performance. This foundation is a core operating system commonly known as Darwin. Darwin integrates a number of technologies, most importantly Mach 3.0, operating-system services based on 4.4BSD (Berkeley Software Distribution), high-performance networking facilities, and support for multiple integrated file systems.

[ Edited by discreetfx on 2009/8/27 21:13 ]

@discreetfx,

Do you really get that much support with MorphOS? Compared to AROS which has mailing lists and forums and an IRC channel too?

The FreeBSD guys made their choice with their license which gets BSD software out to the rest of the world. The same license made WebKit the best browser engine on the planet - so much so that Apple, Nokia and Google's contributions (which they did NOT need to make) continue to make it the best free AND paid for browser on the planet (by paid for, I mean for Apple you would probably do better to be running a Mac, and for Nokia, the chances are you bought a $500 phone). The same kind of license, ironically, applies to the TCP/IP stack running on MorphOS, many of the included tools and features (including PNG icons).

With AROS on the Efika MX, you need to buy the Efika MX. Isn't that a fair enough price for the support you'd get? The original idea was that the Efika port of MorphOS would be free, and Genesi actually paid up front for this (while it was still MorphOS 1.5), and until it took 2 years to actually get there, was still the plan.

Did your 150 Euro make up for the 2 years without an OS update? :)

Let's be honest; there is a lock-in with MorphOS, which is the hardware. It only runs on discontinued products. We no longer actively support the Pegasos and the Efika will probably lapse except for special cases in a few years. Apple have not made a PowerPC Mac Mini for 3 and a half years. It would be better for MorphOS all-round if it ran on hardware that, while still a lock-in to certain hardware models, was actually purchasable without going to ebay or Craigslist.

Without that hardware available, the number of new users will be absolutely zero, and the market can only decrease.

@Neko

We love AROS and are developing a version of Aladdin 4D 6.0 for it. It's amazing how far AROS has come recently. Having a virtual machine version was a great idea to get more people trying it out. AROS should have a basic version for free and a higher end version that costs money. It would bring in more resources to advance AROS even further. You bought up Webkit and it's awesome. Free works in that case since all web browsers have to be free thanks to Microsoft and the killing of Netscape. We don't mind Mac's and our main system is a MacBook Pro with an iPhone 3GS. MorphOS 2.3 is a fantastic operating system, it's future is up to the MorphOS Team. Quality software development always takes time and hard work. Only the MorphOS Team knows what a far price is for their labor.

[ Edited by discreetfx on 2009/8/28 1:06 ]

[ Edited by discreetfx on 2009/8/28 1:10 ]

[ Edited by discreetfx on 2009/8/28 1:13 ]

[ Edited by discreetfx on 2009/8/28 1:15 ]

[ Edited by discreetfx on 2009/8/28 1:16 ]

[ Edited by discreetfx on 2009/8/28 1:24 ]

Matt, I'm curious, is this your personal opinion or official Genesi policy?

Matt, I'm curious, is this your personal opinion or official Genesi policy?

I guess it is official, looking at his signature.

Quote:

Sure, as would I, why would you even mention it you daft bugger? :P

Quote:

Sure, but the important difference here (as is what I was trying to tell you in my first post) is that they were trying to offer a complete and compelling product .. too bad they failed, as I'm sure Genesi will do too unless they come up with something that would actually appeal to a broader market.

Quote:

The same to you, next time you might actually want to read the whole post before replying... :P


- CISC

Quote:

.. And nothing to do with underpowered and lowspecced hardware? :)

@Neko

Reading your posts through the thread, I'm not sure what your goal is. The first impression is that you want MorphOS on Efika MX (and ARM architecture, generally speaking). I guess there are more "business style" ways to achieve this, better than pitching PPC being a dead end (which may be true, but your way of using this fact in discussion is not very nice).

The second thought is you are promoting AROS as multi-architecture AmigaOS clone, working (in the future) on Efika MX. But then promoting a product by bashing its competitors just does not work. It is also worth noting, that AROS software base is thin, even compared with MorphOS.

The last alternative is just you are getting some cheap entertainment on this forum, but then I would't sign it with the professional signature.

Quote:

That is capable of having up to 512mb RAM, but has been crippled by only being available with 128mb and having firmware that does not allow it to be upgraded to any other amount of RAM (or so I have been told).

But isn't there one real poblematic issue with MorphOS: What comes after the Mac port? I definitly like the step by step approach and think it was the best decision to use old Apple hardware to sustain hardware for the next time. But once MorpOS for the Mini (and maybe the powerbook and iBook) is out -- what's on scheudule afterwards?
I think there are precisely two choices how to proceed:
Either Keep the ppc path (however that wil be achieved) or switch architecture.
In the 2nd case the alternative is either x86 or ARM. If ARM gets chosen, then the Genesi products seem pretty fine, but so do a lot of other products. The only advantage for Genesi woud be that in earlier times there was a cooperation between the MorphOS-Team and Genesi. I think the outcome was pretty nice...

[ Edited by Zylesea on 2009/8/28 9:41 ]

@ Zylesea

Quote:

While I understand your curiosity as a user, there is zero incentive to discuss in public what may or may not happen years in the future. If the history of the Amiga has taught us anything it is what damage can be caused when available real-world products are being overshadowed by vapourware.

If there was any confirmation that MorphOS will move to another processor architecture in 201x, this will most certainly not increase interest in the existing PowerPC versions but it could potentially convince some people to maybe not get a Mac Mini G4 now and wait for that yummy MorphOS 5.0 that is to be released at some unspecified point in the future.

The whole idea behind the demo mode of MorphOS is to keep people focused on the here and now and enable them to test the operating system and see if it fits their requirements *as-is* *today*. If it is does, you are free to buy a key to disable the time limitation. If it does not, you might want to resell that $99 Efika mainboard to someone else or keep it if the *free* MorphOS demo is sufficient for you. Just do not buy MorphOS based on what you personally hope it will become some day rather than what it is in the present because the developers are not making any promises regarding the future just to get an extra sale.

Quote:

I think I am with you on that generally, but there is also another way to see it: Without a kind of far going road map one might think (note the conditional!) MorphOS would be on a dead end and therefore leave the train and neither use MorphOS 2.x nor wait for MorphOS 5.0...
Just for the record: *I* don't think that, but others might.

|||

[ Edited by Velcro_SP 24.04.2011 - 07:22 ]

Quote:

What's wrong with keeping a modest roadmap somewhere public? I don't mean exagerations like "MorphOS 3.0 will support SMP, OpenCL, <name your wishlist items here>, etc". I refer to the current PowerPC situation -you know, no new available hardware etc. I wish it was different, but it isnt'. The thing is that sooner or later a decision has to be made, regarding the platform. The MorphOS developers have of course the final word and its their decision, but this secrecy isnt' really doing any good. I don't think they would make a bad choice if they stated eg. a 2-year roadmap of what they think should happen. Heck, they could even discard any comments of their userbase if they wish regarding the roadmap, but at least everyone would know that there is a specific roadmap to be followed. Deviations might happen, but that happens even in huge corporations anyway. It's not a military weapon, it's an OS, ffs! And many people -incl. myself- would see this as some first sign of openness on the part of the creators. Even better, if users (ie registered users, so that it becomes an incentive to get a license) would actually have a say in the decisions, via some kind of poll.

My 2c.

Quote:
Of course it doesn't. It didn't "need" to implement it in the beginning either.

If you haven't noticed, there are lots of really small (as in "userbase") X86 operating systems done from (relatively) "clean table, at least MenuetOS and SkyOS come to my mind immediately, yet I don't think they are THAT popular. And I wonder what would have made this "68k-incompatible" MorphOS on (say) X86 any more popular?

Quote:
Are you trying to claim that supporting (only) PPC somehow makes it impossible for MorphOS to extend beyond / get rid of old Amiga legacy? Because that's what it definitely sounds like. As I see it, it's two different issues.

Quote:

No, you have been told that the amount of memory is a *design decision* that has to be made *pre-production*. The Efika must *be built* with more memory *from the start*. The Efikas out in the wild today was designed and built with 128MB, and no firmware upgrade will change that, it's a HW thing. To make an Efika with more memory you will have to build it slightly different from the start. That wouldn't help you, me or anyone else with a 128MB Efika.