OS4 moves to x86. What happens next?

[runequester]

So this comes up constantly. "Amiga should move to x86 and everything would be way better".

I don't think its all that simple really. Here's some things that'd at least warrant exploring:

1: Time and money required to port the operating system. I have absolutely zero idea how long this would take, but I am assuming it'd be a pretty significant endeavour.

2: Time and money required to port applications.Without app's, an OS is worthless.

3: Oh hey there AROS! What functionality does os4 give that AROS doesn't ?

4: Most overlooked:
Hardware support.
AROS has been around for years and still supports a fairly limited range of hardware.
When it comes to hardware there's two options: The linux way (support it all in the kernel, which requires massive amounts of work from a large number of people) or the windows way (have the hardware manufacturer write drivers for you).
OS4 would have neither.

The assumption in these threads tends to be "we could run amiga OS on any PC and it'd be rad". And that would rad, but it won't be reality.

5: User base.
Is there any actual user base in a world saturated with mature OS choices ? What do I get out of amiga os in the present day and age, that isn't already served by one of the big four (windows, linux, bsd or mac) in some flavour or form ?
I'd love to see data on this, but I heavily suspect that the amount of non-amiga folks using morph or aros is a distinct minority.
I have a hard time seeing how x86 amiga os would attract new people rather than just shuffle existing users around.


Pessimistic ? Maybe. But I think the choices that were made (powerPC) were good choices at the time. I think it's too late at this point to go back.

[Zac67]

So this comes up constantly. "Amiga should move to x86 and everything would be way better".

I don't think its all that simple really. Here's some things that'd at least warrant exploring:

1: Time and money required to port the operating system. I have absolutely zero idea how long this would take, but I am assuming it'd be a pretty significant endeavour.

Considering it's been ported to PPC 'just recently' (in Amiga terms), the amount of work required for x86 should be managable.

2: Time and money required to port applications.Without app's, an OS is worthless.

No need porting apps, a compatibility layer (which can probably be extracted from UAE) will do the trick.

3: Oh hey there AROS! What functionality does os4 give that AROS doesn't ?

Good question.

4: Most overlooked:
Hardware support.
AROS has been around for years and still supports a fairly limited range of hardware.
When it comes to hardware there's two options: The linux way (support it all in the kernel, which requires massive amounts of work from a large number of people) or the windows way (have the hardware manufacturer write drivers for you).
OS4 would have neither.

The driver problem exists regardless of the CPU platform.

[Heiroglyph]

1: Time and money required to port the operating system. I have absolutely zero idea how long this would take, but I am assuming it'd be a pretty significant endeavour.

If they really did convert the OS to C as I've heard, it's not that big of a step.

I've helped completely rewrite a highly complex ~16 million line codebase with just a couple of other developers within a year.

If they can't handle porting a small (at least partially C) OS to a new CPU in that time frame or less, then they need new developers.

2: Time and money required to port applications.Without app's, an OS is worthless.

I don't think OS4's strength is in apps to begin with, the developer community is pretty small.

Most of the apps I've seen for OS4 are Linux ports anyway, so recompile, done.

3: Oh hey there AROS! What functionality does os4 give that AROS doesn't ?

Not so much a functionality, but rather validity.  They own the name AmigaOS.

4: Most overlooked:
Hardware support.
AROS has been around for years and still supports a fairly limited range of hardware.
When it comes to hardware there's two options: The linux way (support it all in the kernel, which requires massive amounts of work from a large number of people) or the windows way (have the hardware manufacturer write drivers for you).
OS4 would have neither.

Aros is a hobby project that had to start from scratch and reverse engineer.  Please don't confuse that with a company who had the source code.

In that light, Aros has moved much more quickly than the official OS.

Even supporting a small subset of x86 hardware, such as one or two chipsets, video cards, etc would still give users many times the hardware availability of custom PPC based solutions at a fraction of the cost.

5: User base.
Is there any actual user base in a world saturated with mature OS choices ? What do I get out of amiga os in the present day and age, that isn't already served by one of the big four (windows, linux, bsd or mac) in some flavour or form ?

Is that an argument against PPC as well?

I have a hard time seeing how x86 amiga os would attract new people rather than just shuffle existing users around.

With the current OS4 hardware requirements, how do you expect to grow the user base?

Nobody "just tries out" an OS that requires a hardware purchase equivalent to two Windows PC's or a Mac.

Pessimistic ? Maybe. But I think the choices that were made (powerPC) were good choices at the time. I think it's too late at this point to go back.

That shouldn't prevent us from going forward though.

In the scheme of things, PPC was just a speed bump.  Most users never saw, much less owned one so the loss in compatibility would be bearable.

[Heinz]

So this comes up constantly. "Amiga should move to x86 and everything would be way better".

I don't think its all that simple really. Here's some things that'd at least warrant exploring:

1: Time and money required to port the operating system. I have absolutely zero idea how long this would take, but I am assuming it'd be a pretty significant endeavour.

They could use AROS as the foundation and simply compile Reaction and other OS4 only stuff, that AROS is missing add a fancy OS4 Theme and ta da ! Theres OS4 x86

2: Time and money required to port applications.Without app's, an OS is worthless.

AROS is API compatible. Most C programs can be recompiled with minor changes.
I dont think there are much PPC Assembler apps for OS4.

Integrated UAE can provide 68k program compatability. (almost done - Janus UAE)

3: Oh hey there AROS! What functionality does os4 give that AROS doesn't ?

The Name. Not much more.

4: Most overlooked:
Hardware support. .

Write drivers for one single Chipset + GPU and you have plenty of Hardware to choose from.

But IMHO it really doesnt matter anymore if Hyperion goes x86 or not. They wont gain any significant market share. They might survive if they find a cheap entry Hardware. That could be a nice ARM Netbook or if they are in Luck, someone produces a PPC based Netbook.

[SamuraiCrow]

AROS is API compatible. Most C programs can be recompiled with minor changes.
I dont think there are much PPC Assembler apps for OS4.

Integrated UAE can provide 68k program compatability. (almost done - Janus UAE)

Writing in Assembly is deprecated on AmigaOS 4.x and as a result, there are no low-level includes for the AmigaOS 4.x SDK.

As for integrated UAE, it's unnecessary if you're running 68k AROS.  

[Amiga_Nut]

Porting the OS to x86 instruction set is one thing. Making it work on the millions of combinations of exact motherboard type, sound chip, video card, firewire/sata/usb devices is the problem.

Hyperion would never have the resources to write enough device drivers to support enough combinations of PC hardware to make it financially viable. People forget that no two PCs are alike and because of that recoding the OS to work on x86 CPUs is only the start. It's a tricky one too because your market is tiny but the number of different combinations of hardware that they need to support is massive. Bad business sense and I don't blame them for ignoring x86 really.

If you ask me they would be better off modifying OS4 so it runs on most recent Macs and provide a nice boing ball sticker to cover the Apple logo

[the_leander]

After OS3.9 we had Amithlon, which was and to this day remains a great achievement technically. If we were to have gone to x86 it would have had to have been then.

Sure you could do a port of OS4 and make an x86 native version I suppose. But what would it realistically offer that wouldn't be better supported with a lightweight linux running UAE?

And this is where we're at now. Sure, PPC, it's cute and at the time it might have been justifiable. But it got the community nowhere.

I have to say that at this late stage in the game. The only realistic options left on the table are UAE and classic68k / Minimig.

Natami might become one if and when it's released.

I just don't think there is the resource there now to do anything else except fall back to the pure retro scene. AmigaNG, regardless of which side of the red/blue divide you're on, has been an utter disaster for the community as a whole.

[Nostalgiac]

Leander is mostly right I think.

Just use/integrate UAE for classic support and...
As going/staying ppc requires very specific and expensive hardware, I don't see why x86 could not be supported in a similar way ?
I mean... just test/release for 1 to 2 mainboards and perhaps 1-2 cpu's.... write drivers for those, and specifically say other ones are not supported. I admit you would need to repeat this each years... but it seems doable to me.

just a thought
Tom UK

[Heiroglyph]

Well said Leander.

IMHO, both of the PPC OS's are Phase5 compatibles anyway.  Commodore and the original Amiga engineers never mentioned PPC as a viable option far as I know.

An Amithlon clone that slowly merged with Aros would be the best way to go assuming you don't have the OS source code.

[Heiroglyph]

Leander is mostly right I think.

Just use/integrate UAE for classic support and...
As going/staying ppc requires very specific and expensive hardware, I don't see why x86 could not be supported in a similar way ?
I mean... just test/release for 1 to 2 mainboards and perhaps 1-2 cpu's.... write drivers for those, and specifically say other ones are not supported. I admit you would need to repeat this each years... but it seems doable to me.

just a thought
Tom UK

Specifying even one exact x86 hardware configuration is still easier for users than specifying four proprietary PPC configurations.

That's all Apple does with x86!

Plus, I guarantee finding a specific PC motherboard even five years after being discontinued is cheaper and easier than finding the custom PPC today.

[orb85750]

I don't see the average x86 user wanting to pay for AmigaOS unless it's much superior to others.  What will AmigaOS get people on x86 that they cannot get for free with other OS?  So it's really "just" the current Amiga users that will find the easier route (x86) to using OS4?  Is that the main point here?

[the_leander]

Specifying even one exact x86 hardware configuration is still easier for users than specifying four proprietary PPC configurations.

I always figured the best way to go about it would be to pick something like the mini-itx systems by VIA and more recently Intel/nVidia.

Limited hardware differences, all well documented and quite a few cool features.

As for what they'd offer over other things in the market, that really is the $64,000 question, isn't it?

[Heiroglyph]

I don't see the average x86 user wanting to pay for AmigaOS unless it's much superior to others.  What will AmigaOS get people on x86 that they cannot get for free with other OS?  So it's really "just" the current Amiga users that will find the easier route (x86) to using OS4?  Is that the main point here?

At the current state of any Amiga-alike, there is no reason to choose it over a "standard" OS.  It's stagnant and split too many directions.

But, if you are competing with only Amiga-alikes, then IMHO an x86 based OS could easily gain the majority of the users and help unify the community.

Honestly, even hard-core Amiga fanatics are balking at the price of entry on PPC solutions.

[persia]

Yes, I don't think any AmigaOS will ever overtake Linux in terms of popularity.  The reasons behind the idea of X86 AmigaOS are simple.  X86 equipment is plentiful and cheap.  An X86 Amiga would allow you to load a Virtualbox MS Windows/Linux?MacOS system to run software that doesn't exist on the Amiga.  If Wine could be ported you could run many MS Windows programs outside of a virtualbox.

Basically for the price of AmigaOS you could have a brand new Amiga.

[lsmart]

I mean... just test/release for 1 to 2 mainboards and perhaps 1-2 cpu's....

The mainboards would be outdated and unavailable before the OS is even released. Seriously, if you are into this kind of stuff - look at AROS. It won´t get any better than this if you are on x86. We don´t need a second AROS, we already have one. The only way you could improve an x86 Amiga-like experience you have to nail the API on top of Linux which has no driver problems. But according to some MorphOS fans on this site, Anubis isn´t going well at the moment.