OSDev.org

Maybe one of us here can help each other think up something to help revive them.

Here are two mine: Project Xanadu and the Lisp family of languages. Sadly, I couldn't find any videos on my third, synthesizing kernels, so I'll just leave this link instead.





Actually, Lisp isn't quite so lost now with Clojure gaining ground, but I might as well include it - this isn't the first time it looked like it would make a comeback.

In an OS development forum, it would have to be OS/2.

AmigaOS x86_64

An intel/amd based, with modern GPU and sound support, of that classic and revolutionary operating system!

Yes, I have a great nostalgia for AmigaDOS, and also for Sinclair QDOS.

iansjack wrote:Yes, I have a great nostalgia for AmigaDOS, and also for Sinclair QDOS.

Oh well, I can always fire up WinUAE and run it fullscreen...
I miss the good old day too. Like The8bitGuy on youtube said, what does your new computer do that your previous one didn't. Compared that to when you went from your Sinclair to your Amiga for example...

I think the change from 8-bit computers to the QL (with it's strange 8/32-bit CPU) was probably greater than that to the Amiga (although that was also a quantum leap - pun intended for those who remember the adverts). I didn't have to keep sending the Amiga back for a new one; it actually worked!

iansjack wrote:I think the change from 8-bit computers to the QL (with it's strange 8/32-bit CPU) was probably greater than that to the Amiga (although that was also a quantum leap - pun intended for those who remember the adverts). I didn't have to keep sending the Amiga back for a new one; it actually worked!

Yes Sinclair (and later Amstrad) were known for cutting corners when it came to quality control.
But, since I'm Canadian, I never had the pleasure of using any of their computer. Only the ZX-81 under the Timex rebrand made it here.
The big 3 here were Radio Shack TRS serie, the Apple, Atari and the Commodore brand machine. There was also the Texas Instrument TI-99/4a but it was underpowered and overpriced compared to the others.

I started on a TRS-80 MC10 and moved on from there. Each new one bringing a whole new world of possibilities and features. Today, all that we get when replacing our computers are a slight speed bump and a few less expansion slots and ports...

AMenard wrote:Texas Instrument TI-99/4a

I had one of those!! I was happy because the keyboard had REAL key where as the 99/4 (without the a) had calculator keys for the keyboard.

eryjus wrote:

AMenard wrote:Texas Instrument TI-99/4a

I had one of those!! I was happy because the keyboard had REAL key where as the 99/4 (without the a) had calculator keys for the keyboard.

I loved its modular expansion box... It did look badass for the time.

Dumb terminals. Serial ports.



Beautiful and simple. (OK, I admit I would prefer a more "modern" design, but the concept is beautiful and simple.)

I'm seriously thinking about making a dumb terminal with an Arduino and a bunch of 16-segment displays (or maybe a dot-matrix) one day. And, of course, making a driver for it for Linux and my OS.

computersandcoffee wrote:Dumb terminals.

IBM 5250 emulation is alive and well.

http://tn5250j.sourceforge.net/

iansjack wrote:In an OS development forum, it would have to be OS/2.

I'll second that, and add BeOS.

I still mostly use OS/2 / eComStation, and I do have BeOS / Haiku in a few virtual machines.

In fact, my OS development is being done with eCS 1.2 as the host operating system, EPM as the source code editor, and Watcom compilers, assembler and other tools.

I've started looking at the Mill architecture recently, and while I need to do a lot more research on it, from what I've seen so far, I think I'm in love with it. It is very similar in some ways to my RSVP idea, but with actual understanding of the technical issues - it solves most of the problems I found in the RSVP idea, and many I didn't notice before now, and in a way that (unlike RSVP) actually is feasible and potentially highly performant.

The belt machine design concept in particular intrigues me, as it seems to have most of the advantages of a stack machine while avoiding or mitigating most of the disadvantages. Hell, Geri should take a look at it, because it might be the thing to make a modified form of SubLeq a feasible architecture (though the actual Mill design goes in the other direction with a packed VLIW approach).

However... I have real doubts if it will go anywhere. Even the ones working on it agree. One of the designers gave a series of lectures on the design in 2013, and in one of them, he explained just how big the hurdles it - or any new architecture entering the commercial market - faces in the first 2 minutes of the video.


It is too early to call it a lost cause, but it will require a lot of work, money, politicking, marketing, and most of all, luck, for it to even reach completion, never mind any significant degree of success.

Linux on the desktop.
I love the idea of completely replacing MS with free software but Linux (at least Debian style distos) have the worst user experience imaginable and it doesn't appear to be getting much better.

StudlyCaps wrote:Linux on the desktop.
I love the idea of completely replacing MS with free software but Linux (at least Debian style distos) have the worst user experience imaginable and it doesn't appear to be getting much better.

Ubuntu was going in the right direction for a while but it seems to have relapsed a little bit (I still use it approx. 90% of the time, though)