Now this topic isn't talking about value for money or anything like that as far the x1000 goes. I was watching a video recently linked on here that talks about the new CPUs to be used and they mentioned that they will still have the Xmos on the new boards just like the x1000. Now I may not be up on the whole scene for the x1000 and what has happened to it but I don't remember hearing that anything has been done with the XMOS chip on all these machines.
There are things going on - but nothing definite released yet as far as I know.
My question is simple, is it not possible to program the Xmos chip to replicate the functions of Agnus/Denise/Paula (and the AGA equivalents) in the FPGA to essentially remove the requirement for UAE? Think of it as on-board minimig style board as what I am trying to explain.
Which FPGA are you talking about, exactly? I don't think there is an FPGA on the X1000 is there? I know there is on the Sam boards, but not the X1000 I don't think.
You will always be tied to UAE, though, I'm afraid - there's no escaping that. The xmos could possibly be used for AGA-like functions, but it'd be of very little use because there'd be no way of using them for games (95% of which are hard-coded to use the chipset).
I think what you're wanting is to be able to boot off a game disk or maybe run WHDLoad and have it run natively.... but while that'd be rather cool it's just not possible without an emulator, as the Amiga games are too closely tied to the hardware.
I can understand that floppy drives are no longer being mass produced so there is no point talking about a replacement disk controller. I think if we could get a NG Amiga that essentially has firmware hooks to run old code that requires OCS/ECS/AGA and build it around a budget CPU (single core, given that OS4 is a single core based OS).
OS4 is in the process of migrating from single-core.
The price of the CPU is only part of the problem. Take the X1000 - it costs around £1400 + VAT for the board, of which about £600 I believe is spent on the very expensive CPU. That still leaves £800. Nobody's getting rich off the X1000 - that money is because of development costs. Even if you put a dirt-cheap CPU on a board, the development of that board is what costs the money.
I think many people would consider buying such a machine given the OTT prices for some big box Amigas and it would remove a barrier that make many people feel NG Amiga is not really an Amiga at all given they use the same technique to run Rocket Ranger as a Windows PC or OS X Mac (ie UAE emulation).
If by the big box Amigas you mean the AmigaOne and Sam machines - they're really not OTT at all - you have to factor in all the costs of development. When you mass-produce goods the development costs per unit are tiny, but on a machine you're only producing a few hundred of, they're massive (hence the cost per unit to the buyer). In fact, when you consider the usual prices of boards custom made like the X1000 board was, it's an absolute bargain (you can easily pay 10,000 Euros for a prototype board in the embedded industry). Of course you can't just decide to mass-produce, either, because that requires a massive outlay of cash, and you may well supply more than there is demand for.
The biggest problem I think, though, would be you'd end up with a machine that didn't know whether it was a next-gen Amiga or a Classic one. The next-gen element would be compromised by the use of the single-core CPU, and the classic element would never run as well as a real Amiga, and would always cost much more.
Also, even if it did run classic Amiga software natively, I think the die-hard Amiga users would still hate it anyway. For many of them, if it doesn't say "Commodore" on the box, it's not a real Amiga.
I do, however, take your point that it would be a good thing to be backward compatible. I think the best solution for this would be a Minimig-on-a-Card. Stick it in a PCI-E slot, and have it work like the bridgeboards do on the classic Amigas - let it share graphics (a simple electronic monitor switcher would do), and take the keyboard and mouse input from the host like janus.library does. Even better, much of the software development is already done by the Minimig guys (though I'm not sure what the license is). That's what I would look at, anyway.
