Why do so many people assume its such a lengthy and expensive process to port an OS to another architecture?
Ive seen figures in the millions mentioned more than a few times, and timeframes measured in years.

AROS is a good example to show this simply isnt true. Current Sam460 build took only a few months, and by one or two people in thier spare time. Sam440 version took about 6 months by one person, again in their spare time (and they wrote new drivers to go along with it).

Id be very surprised to hear of even a 5 figure some was required.

As for the original question, accessibility. This means both more affordable (improved specs would be nice, but less important), and not being little more than industry standard hardware, with a few exotic things that make the price out of reach (while adding nothing vs. other options that are easily sourced/cheaper).

The current computer climate simply doesnt allow for such bad price/performance. There's way too many cheap, more functional options out there for OS4.x to be anything more than a hobby for a few people.

Some of us might swallow it, but its not going to see more than a handful of new/returning users.

Last edited by fishy_fis on 22-Oct-2012 at 03:05 PM.
Last edited by fishy_fis on 22-Oct-2012 at 03:04 PM.

@fishy_fis

"AROS is a good example to show this simply isnt true. Current Sam460 build took only a few months,"

Based on SAM440ep codes I believe.
Doubt it already supports everything? Being compareable to AOS4? DMA? Warp3D? And all AROS apps compile for Aros PPC ? ...

(hmm... I should be using AROS + nouveu + Gallium3D on my SAM440ep already... )

" and by one or two people in thier spare time. Sam440 version took about 6 months by one person, again in their spare time (and they wrote new drivers to go along with it)."

I think it is very simple to target to only one set of chips. There is no such target on x86 without a custom board.

(Perhaps Hyperion would need to borrow some miracle workers from AROS team)


I know Hyperion guys have said it is simple to get AOS4 to another architecture (a few months amount of work) but I think that would be only wery quick and dirty port. It has taken 10 years while trying to reach AOS3 maturity for AOS4 on PPC.

Last edited by KimmoK on 22-Oct-2012 at 05:49 PM.

_________________
- KimmoK
// For freedom, for honor, for AMIGA
//
// Thing that I should find more time for: CC64 - 64bit Community Computer?

@KimmoK

Hyperion has failed to properly support their own custom boards so that is a pretty invalid argument. At least with the x86 you would have the brute force of the CPU to help while waiting a few years for the drivers.

_________________

@Petah
I am not entirely sure what you are on about but next time I suggest you ask me what I meant before charging off next time.

_________________
ExecSG Team Lead

I too have always wondered why people say it would take years and years, and vast sums of money to port to a different architecture. It may take a fortune and years to get AOS up to modern day standards, but for a simple port... not sure why it'd be so expensive.

And yeah, when people say x86 port, they don't necessarily mean to every single PC variant in the world (with associated drivers). A custom x86 or ARM system would be much, much cheaper than the $3000 PPC machine they have now.

One thing I do wonder about is in regard to desktop usage. Sure, AOS can be used as a desktop OS. But based on what I have read here (and I may be wrong), AOS is like 10 years behind mainstream operating systems? I was sort of baffled when I read someone here comparing it to Mac OS 9.22 in a different thread. Unless they meant AOS 3 vs Mac 9.22, which would make more sense to me. It's not that far behind, is it?

But if they want 50K users, why assume they have to go for desktop users? If the OS isn't really up to date as a desktop operating system, why not go for an internet device/cheap gaming/multimedia appliance route instead? If they had money years ago, I would have thought going the Roku/Google TV route with a general Amiga ARM internet/streaming device would have made sense. Why not try to gain some users (and money) and then try to go after desktop users after they have a decent size userbase? Then they could use those resources to beef up AOS so it's at least competitive.

@KimmoK

The point being that if 1 or 2 people can port to a new target in reasonably short space of time in thier spare time then some people are clearly over-estimating things.

Also, its about the OS, not the apps or 3rd party additions, and Id be very surprised if people developing it arent making it at least reasonably portable.

As with my AROS example (something you yourself went on to support) supporting a new platform is a far cry from writing from scratch, even with a new architecture.

This is all a little off topic anyway. I just couldnt refrain from commenting on the often heard/repeated myth.

In regards to the original question, the only way it'll happen is if there's a different plan of attack to what's being taken now. Exactly what that plan is I have no idea, but something has to change. To even become a niche system (as opposed to a niche within a niche within a niche) it has to be more accessible. What's been served up thus far just isnt condusive to attracting new people. Not a criticism, just so very obvious. Id be surprised if there's many people that disagreed there.

At the end of the day Im not too bothered. I enjoy my amiga hobby and probably always will, but I honestly believe OS4.x is on a bit of a self destruct course unless things change. You simply cant survive commercially these days being tied to the sorts of expensive systems its currently tied to.

@Frags

Quote:

Frags wrote: @Franko Just google `hackintosh` or `osx86` and there`s a raft of info out there

Well after googling "hackintosh" and spending all night reading everything I could find on the subject, I have given up on that idea and have just gone ahead and purchased a brand new 27" iMac i5 thingy...

Strange, I can tinker away forever on my old miggies getting things to work (or not) but somehow after reading all things "hackintosh" it just didn't have the same appeal for me to "tinker" around and build my own system for OS X... must be an Amiga thing or something to with with old technology but doing it on modern hardware just didn't sound like fun...

_________________

@Franko

Too late now, but... There are bootable CD images that one can download and with the proper hardware specs, it is as easy as booting first with the boot CD, then inserting the Mac OS install CD and following along as if one were installing on an actual Mac---you do have to format your drive to HFS, of course, during the system install process by using the disk utility.

http://wiki.osx86project.org/wiki/index.php/Installation_Guides

I used the myHack CD. One should look for the the CD image with the most drivers and automation (autodetection).

Please excuse this little interruption of the thread.

@fishy_fis

"OS4.x is on a bit of a self destruct course"

It might be.

"You simply cant survive commercially these days being tied to the sorts of expensive systems its currently tied to."

I still fail to see very cheap alternative on the market.

But I imagine that also for low cost PPC board, AROS support is far more easy to get than those commercial ones.

If I had opensource PPC board betatester boards ready and I sent some of them to AROS developers, I would not be surpriced if I get back a working AROS binary pretty soon... Hoping to be able to test that scenario one day.

_________________
- KimmoK
// For freedom, for honor, for AMIGA
//
// Thing that I should find more time for: CC64 - 64bit Community Computer?

@azhoward

Something to try, the next time I have a spare x86 HW.

_________________
- KimmoK
// For freedom, for honor, for AMIGA
//
// Thing that I should find more time for: CC64 - 64bit Community Computer?

@KimmoK
The cheap alternative is an x86 system build by Amiga dealers. Exactly like Vesalia does with AROS systems.

Quote:

But I imagine that also for low cost PPC board.....

There is no low cost PPC board and never will be as there is zero demand outside the AmigaOS community.

-----
To gain more users I think we need to unify the Amiga NG systems (at least 2 of them, which would probably leave the third way behind).
Why? Simply because the OS itself is not really nterestng to outsiders yet and development speed is WAY to slow when the same work has to be done on multiple Amiga platforms.

If was in charge of MOS or AmigaOS I'd be looking intense at how a merger with AROS could be done. If you need to take a step or two back to go in the right direction just do it.

_________________

The people in my area of the world seem to think that ther are only two types of PC's.

Based on that statement, ther are two things that we should focus on in my opinion.

Software!!!! Specificaly, web browsing and office suite.

Advertizing - You can have the greatest product out there, but if nobody knows about it, then it's all for nothing!

@Petah

I voted the last option as felt something was missing. Release that damn netbook for $300 bucks !

A cheap laptop like Amiga system. That's gotta sell. But they gotta get it out before it's too late!

x86 port is easy to pick but that wouldn't be fair, as it would just end up being pirated, and soon enough unlike today you'd see torrents and illegal downloads for the updates.

@Franko

Right now we shouldn't need 68K compatibility anymore. All the apps we need should be on PPC. Like how the Mac was ten years after PPC was introduced. It does seem logical to port to a new architecture with a clean break and introduce 64-bit and SMP.

But here we are 2012 and still with a desire to run 68K software and games. What's funny is that we shouldn't need to as some old software like the native chipset. So although now we use VGA it still isn't good enough to replace AGA. Once we hated bit planes and CLUT only modes but now we want it back so HAM8 works along with copperlists! Even if OS4 used VGA raster interrupts which is the closest to what the copper could do it still wouldn't be good enough!

Funny to think that AGA held us back and we wanted VGA but now VGA is holding us back! Still without some kind of copper it's hard to reimplement AGA modes again. We should be over them but Amiga CLUT modes could change a whole screen with one palette register change. Even if modern 32-bit VGA is faster it still needs to fill the whole framebuffer to do the same by whatever means. Still looks inefficient by comparison. Perhaps why we can't move on from AGA.

@KimmoK

Quote:

1. It's way too easy to throw in x86 ARM suggestions. There are no dektop capable ARM chips yet. (perhaps next year, perhaps not) 2. There is no way to create stable system on x86 without custom board. (price will not be cheaper than with PPC) 3. AOS4+SW port to ARM or x86 will cost something like 100 000+ eur or one decade

2. If you begin porting AmigaOS to x86 today, you have to make a plan, trace a roadmap and decide a deadline. Then, you'll have to decide a hardware target that will be a viable, not obsolete system at the time of that deadline. For instance, if you decide to build a x86 amiga system due to be out on december, 2013, you have to create something what will be viable for december, 2013. The difference with PPC, is that you may count on dozens of big players (ASUS, MSI, ASrock, Sapphire, Zotac, GigaByte, and more more more) who already have their middle and long-term plans, you can sign a NDA agreement with them and eventually buy a specific motherboard model in a certain quantity. It's far cheaper than doing everything by yourself.

1. And this is why I answer this after that: the same plan applies to Google's partners and anyone making ARM processors and embedded systems.

3. I don't think AOS sources are so badly written they can't be ported to a different architecture! But seeing things from the other side, you should really thak the AROS team for their effort: after all, they recreated everything from scratch, making a much greater effort than just adapting and recompiling things... so they made you a far more valuable than Eur 100,000 gift.

I've been advocating "AmigaOS on x86" for many years, and I still believe it's a viable path to ensure long term survival of this wonderful OS. Maybe the only one that actually makes sense, I'd argue.

Having extremely cheap HW would be nice too, but nowadays I'm a bit short on space than I used to be during the old good days, so I'd prefer to be able to use commodity HW and a fully-software AmigaOS offer.

I do see the point of people (and companies...) wanting to keep proprietary hardware, it's just that I don't see that as an added value anymore. Definitely not when proprietary hardware is so extremely expensive and not as cutting edge as it was before Commodore went bust.

Last edited by Turrican3 on 23-Oct-2012 at 03:29 PM.
Last edited by Turrican3 on 23-Oct-2012 at 03:26 PM.

General question for those who have used AOS:

How much more advanced is AOS 4 vs AROS? Does it have a lot more software for it? Is it much nicer and have a ton more features than AROS?

I was just wondering about what may happen if AOS did go the x86 route. My first instinct was that it'd get a lot more users. Then I began to realize that AROS is already x86, free, and I doubt it has a userbase of 50K users.

So would it even matter if AOS went x86? If it was given away for free right now, and worked on most x86 systems out there, would it get 50K users who used it regularly as their main OS?

@Troels

>The cheap alternative is an x86 system build by Amiga dealers.

It still costs some 400eur minimum.

>>But I imagine that also for low cost PPC board.....
>There is no low cost PPC board and never will be as there is zero demand outside the AmigaOS community.

Was there any demand for the pathetic Rasberry Pi before they made it?
No. And soonish they have sold those 50k units.

_________________
- KimmoK
// For freedom, for honor, for AMIGA
//
// Thing that I should find more time for: CC64 - 64bit Community Computer?

@paolone

"you can sign a NDA agreement with them and eventually buy a specific motherboard model in a certain quantity. It's far cheaper than doing everything by yourself."

Sounds too simple...

Anyway, whatever the HW is, as long the systems gives me AmigaOS fun (in nicer/cheaper way than today), I'm all for it.


(in the meanwhile ... as long as most amigaos variants run on PPC, I search for the way toward 200eur PPC HW)

Last edited by KimmoK on 23-Oct-2012 at 04:44 PM.

_________________
- KimmoK
// For freedom, for honor, for AMIGA
//
// Thing that I should find more time for: CC64 - 64bit Community Computer?

@Petah

port it to 68k and make sure it runs on winuae. Let Claonto distribute it.

_________________
Support AROS sponsor a developer.

"AROS is prolly illegal ~ Evert Carton" intentionally quoted out of context for dramatic effect