Hans_ wrote:
It was called BeOS.
Worked a damn sight better then OS4 ever did too. But going back, it is unfeasable, because simply "creating a new API" with the sorts of fundamental changes you would need, with the current limitations means you would basically have to re-write the thing from scratch, whilst at the same time breaking compatability with all previous software, at which point, why bother? There are off the shelf solutions out there that would probably be better suited, hell, Haiku - the opensource BeOS clone with a UAE sandbox integrated would probably give you everything you needed and give it to you in an identifiably AmigaOS like way. Aros is another potential way forward...
It's not as insurmountable a problem as you make it out to be. Obviously it'll take effort and would take time. The new API could of-course draw from existing solutions. In fact, I'd expect them to use industry/de-facto standards wherever possible.
Do you have any concept of how difficult it is to produce a modern operating system that meets todays standards?
Do you?
Now, I was once an Editor for Zetanews and as such had access to developers in a way most people don't, I talked with them in detail discussing both Zeta and later Haiku as they built up from a far better starting position then AmigaOS offers.
Haiku is probably the best place to start in terms of just how complex the task of reimplimenting a modern OS from scratch, they had some truly brilliant people working for them full time and it's still taken them 7 years to get to the point they are now, and what they have still isn't complete or as capable as BeOS R5.03, released back in 2000!
AmigaOS reimplimented in a modern fashion would be and is a pipe dream, one lacking in both resource and a framework. To take the AmigaOS and make it a truly modern OS, with all the things that are expected from a modern OS (SMP, multi user, memory protection etc etc etc) would be a staggaringly complex proposition, far greater then "simply" reimplimenting what came before (such as AROS or Haiku), requiring far more developer numbers then even both communities combined could offer.
And that's ignoring the complete lack of a roadmap to work towards... And no, I would seriously not want to be the one to open that particular can of worms.
Hans_ wrote:
Why do it? To push the system forward. We all know that breaking compatibility is required so why not just go ahead and do it?
Yes, great, but at that point, why bother? Seriously, if you are going to drop support for old applications, why not simply move to a new OS and have done with it?
You might as well take a snapshot of AROS and Haiku code, shake em together and slap an Amiga sticker on it, it'll save you time and effort and the net result would have about as much to do with the Amiga as any from scratch OS you came up with.
Hans_ wrote:
Why make the NatAmi? It's never going to compete with mainstream computers; it's graphics features are never going to compete with the latest graphics cards. No existing software will make use of its advanced features.
I'd say in both cases it should be done because people enjoy it. If I were involved, I'd probably get a kick out of simply proving people like you wrong. The Natami team are doing stuff that other people say are unfeasible.
There is nothing unfeasible about NatAmi, it's an FPGA board thats been told it's an amiga-alike. Now the cool factor there is the same as the C=One, it also offers the potential for developers to cut their teeth on relatively simple and elegant solutions on which to build better things, the ultimate bedroom coders board. Truthfully I feel that something like this or Amithlon *should* have been chosen as the way forward for the Amiga.
The NatAmi project holds the potential to be of interest to far more people then any number of AmigaOne models simply because of the nature of FPGAs.
Getting bored with your "Amiga", select the AtariST/Falcon rom image and hey presto! Instant Atari, or whatever.
As I said, the scope for Natami in this regard is great, and not just by this community.
Hans_ wrote:
That's the best that you could hope for, not everyone else. There is no technical reason why Amiga OS could not utilise something more complex or capable.
I don't deal in false hope, it always ends in bitter dissapointment. Reality is my prefered choice, and the reality is that the best you really could hope for is something like an Amithlon/Aros hybrid running on X86, or maybe NatAmi if you want something really cool.
Your perception of reality is somewhat different from mine.
Hans
I know it is, I have some idea of the monumental difficulties involved in doing what you're suggesting with regard to the AmigaOS.
