> In the world of multicore and GPGPUs there isn't a space for the old Amiga, except due to nostalgia.

For the old Amiga, no, other than nostalgia which is clearly what these guys are aiming for.

But there's still lots to learn from it. I still yearn for a system that consistently remains as responsive to user input as it did, for example (which it did by a meticulous eye to process priorities and decoupling "everything" into separate tasks; e.g. due to potential slow IO, the clipboard was read/written by a separate task; the console/terminal involved about half a dozen tasks etc.). I also still wish a datatypes like system would take hold, so every app could read/write every image format, including ones not year written, for example (25+ years on, and Amiga apps still gets support for new image formats - even the apps we don't have source for).

And on the Amiga, "everyone" used libraries like XPK that made transparent support for whatever compression algorithm someone preferred in any application that used it, for example (and even if they didn't, you could easily add transparent XPK compression support to your filesystem).

And the pervasiveness of AREXX ports. Even though I detest the language, the presence of AREXX ports "everywhere" was a big deal. Even more so because it was a well know features. "Everyone" knew about AREXX, whereas I've yet to see anyone I know who use OS X actually use AppleScript.

And I still "emulate" Amiga screens in the way I use workspaces on my laptop, with most apps maximized (thankfully, on Ubuntu with Unity I finally have a global menu bar again for most apps).

I have an apple script that syncs the itunes library with another computer on the network.

That being said, I think it's usefulness is as limited as you seem to think it is.

> Anyone trying to replicate the Amiga in 2014 doesn't understand what make it so special 20 years ago.

I think those who try yo replicate the Amiga in 2014 and has any rate of success understand that to a greater extent than most of us. I am glad that I have my A1200, but emulation is getting really good, and is more or less the most comprehensive and accessible documentation of the hardware we have.

> In the world of multicore and GPGPUs there isn't a space for the old Amiga, except due to nostalgia.

As someone who didn't really grow up with computers like the C64, VIC-20 and MSX line, I have certainly had a lot of fun developing for these platforms unencumbered by any innate sense of nostalgia. I suggest reading this fun post by Viznut (of PwP fame): http://www.pelulamu.net/viznut/blog/2008-08-21/

Ever since I studied how the *BSD and Linux kernels worked I felt it was not very elegant, it used to be like a web of spinlocks(analogy taken from a mail by Terry Lambert, FreeBSD developer). I think FreeBSD and Linux today use kernel-side threads and mailboxes a bit more, but it's not the foundation of how they work.

I'm hoping for an operating system where the kernel works like the AmigaOS exec.library with Tasks, MsgPorts and Signals, but offer a barrier where the outside could be a traditional ABI like the one Linux offer.

I've not looked, but I think Matt Dillon's DragonflyBSD is heading towards something like that.

Have you looked at Minix yet?

I have the book and have read some of it years ago. I can't remember if the OS-tasks were an isolated process or just a thread in shared memory.

My first comment would be for a kernel built up using threads in shared memory communicating with signals and mailboxes.

Since 3.0, Minix changed the architecture. Now they are full micro-kernel, use NetBSD userspace tools, have EU funding and some commercial clients.

If you wish I can provide more info.

They are separate processes. In fact it is possible to restart or replace then on a running system.

You're right, the Atari ST wasn't on the same level, it was $500 cheaper and had a faster processor.

And still it was slower for almost everything, and had an OS befitting an 8-bit...

EDIT: Now lets get on USENET and start another Amiga vs Atari ST flamewar on comp.sys.aiga.advocacy.

A ZX Spectrum +3 was better that an Amiga and Atari ST in all universes. xD

Don't forget the MIDI though.

Ah, good memories of flamewars in pre-Internet age. The Atari ST was a good piece of hardware, but the Amiga was superior in almost everything it did.

The Amiga was more powerful, thanks to its multiple chips offloading the CPU for many tasks, but the ST was a simpler, more elegant, design. In the end, the evolution of the ST family was easier, as the machine was less tied to NTSC timings and other design compromises that betrayed the gaming console origins of the Amiga. TOS/GEM was a very limited stack, but we shouldn't judge the hardware by its default software.

Thanks for bringing some memories back, I had a 520 ST after my Amstrad CPC664.

That sounds like me. After that, I went OS/2. These days I fear for any platform I choose.

Nostaligia is what this is for, and that's a legitimate reason.

And why wouldn't this work today? Imagine getting an awesome piece of kit like your PS3/4 and being able to code on it, run anyone's code on it. Hack it to your heart's content. I reckon that would (a) rock. (b) foster an amazing community. The things that made the Amiga rock could make any specialist piece of hardware rock, what's prevents it is that companies like Sony and Microsoft are anal control freaks.

I have hope for the raspberry pi to fill a similar niche. The only thing that's held it back so far was the lack of documentation for the GPU.

That box was genius. I am to this day grateful that it, not any other platform, was my entry to programming.