Well, according to the cult of the Amiga, you can't replicate the fluidity and responsiveness of Amiga with Linux. Amiga was about the hardware and software. The hardware was quirky, cool, and cutting edge. To use an Amiga in its day was like a trip to the future. Plus it didn't have any memory protection, so a single goof in a program would kill the entire system sadly you can't do that with just a gui in linux. At least not without creating your gui in the kernel.

>> why Amiga continues to inspire people today

Um...not really. I owned two Amiga machines and worked on two different Video Toaster rigs. Fun at the time, but I'm very, very happy that the Amiga's best features (graphics, sound and text-to-speech) went mainstream. I haven't plugged in any of my old systems in more than five years.

Let it rest - RIP.

Just take a gander at this thread from 3 mnths ago. [amiga.org]

and those are just select quotes from the first 2 pages. The snobs still exist.

That's exactly what they did. Icarus Desktop is a distribution of AROS, in a prepackaged VM image to make it easy to use. AROS is similar to WINE, in that it can run programs within a hosting OS. It also has native ports, but those progress slowly on the hardware support side.

Yeah. What we need is a Linux kernel module that traps userland segmentation faults and throws a kernel panic. XD

I love all my Amigas; I fought on the epic frontlines of the Amiga versus Atari BBS flamewars before most of you were an ache in your daddys' groins. I carried the Boing Ball flag into harm's way too many times to count. But the true Amiga experience, as depicted by connoisseurs, requires abandonment of such niceties as memory protection and process isolation.

The hardcore nostalgics forget that the Amiga didn't have memory protection first because the hardware wasn't routinely available, and more importantly because the seamless memory map allowed all of RAM to be a huge playground for the CPU and custom co-processors to accomplish amazing things at less than 8 Mhz. Also, the kernel was blazing fast because there was no meaningful context transition from userland to kernel; everything was memory-pointer based, and all memory was directly mapped and non-virtual.

Therefore, it was also fragile. But that was an acceptable cost for blazing speed and jaw-dropping media performance at a time that MS-DOS machines were single-tasking, playing beeps and boops through a 2" speaker in the system case. and displaying EGA-level graphics.

So, let's not wax too nostalgic. True nostalgists wouldn't want this any more than an intelligent car collector will settle for a kit car body, even if it's on a more powerful and capable chassis than the original 1950s Ferrari (for instance).

Amiga enthusiasts who are curious or interested in one evolutionary path of the old OS might want to see this.

Other than that, I can't imagine this being a very popular product.

Did you feel a little breeze in your hair? Yeah, that was the comment going over your head. He intentionally switched them up for effect.

Over time, there probably isn't a lot that's better than what we have today, beyond efficiency and a look and feel that I just felt comfortable with - which itself is somewhat subjective... hey, take a look here: http://home.datacomm.ch/mrupp/TAWS/WB.html [datacomm.ch]

At the time however, these were considered radical:

- Pre-emptive multitasking
- A shell that was half way between Command.com and Bourne. Had some very nice user friendly aspects, such as named parameters and a shared command line parsing system.
- The file system supported mixed case, long, filenames.
- An automatically-managed multiple desktop system. Larger apps would have their own desktops. Each could be a different screen mode if necessary (important in the days when there was a resolution/colour tradeoff)

Everything was patchable and extendable. Utilities were encouraged to intercept standard library calls for all kinda of stuff. The file system had some extremely nice features such as an assignments system that allowed you to assign symbolic names to directories - you didn't have drive letters or a single file system, but something more partitionable. From Workbench 2.0 onwards it had an extremely pleasant look and feel (older versions, not so much.)

It's hard really to describe how radical and better it was at the time to anything else mainstream. Unfortunately, it became obsolete the moment MMU support (and other security features) became important, which is to say, as soon as the Internet proper came on the scene.

Take it from someone who used to program Copper every day in the early 90's: It's not exactly a miraculous superchip that is hard to emulate. Games are usually pretty easy, the Amiga demos are harder to emulate 100% thanks to various copper effects.

Sure, Amiga architecture was way different from PC, but these days PCs are so powerful they can emulate 68000+Blitter+Copper easily. I'm pretty sure UAE has this covered pretty damn well by now.

No, the copper isn't, but what the copper triggers, when is somewhat harder.
You don't have a lock to scanline in an emulator. For the display effects, you need to emulate every scanline separately and generate a complete image as it would have looked at the end. No biggie when you have fast enough hardware.

However - and here lies the devil - the copper isn't just used for display effects but can also control a bunch of other hardware registers. (Including audio, where waiting until the end before presenting results won't do what you want - a simple example being creating a 100 or 120 Hz tone by doing two equally spaced copper moves per frame. And the CPU.)
And you can even modify the copper list while it is running, triggered from either the CPU or the copper through a copper-scheduled blit, or both. Then it becomes a small headache to get things done in the right order at the right time.

Getting some of it right is trivial. Getting all of it right is not.