OS Ideas

[ken27238]

I have a Power Mac G4 that I have running 9.2.2, 10.1.5 and Ubuntu 6.10. I am wanting to try a different OS than the ones that I have, I have a Quadra 660av that i can use. but there is a catch, I don't have any of the CD Caddies for the drive so I would have to use floppies

I'm trying to expand my OS experience

[Dog Cow]

Learn C and write your own OS.

[ken27238]

sarcasm

[Gorgonops]

A 660AV is something of an odd duck and therefore used to be a bad candidate for running non-MacOS operating systems on. First thing that comes to mind is NetBSD, and NetBSD's machine status document:

http://www.macbsd.com/macbsd/macbsd-doc ... ne-status/

Says both ethernet and sound are broken on that machine. Meanwhile, OpenBSD says:

The following Macintosh models are supported and tested. This means that at least the SCSI controller, serial console and on-board ethernet will function on these models. On some of these machines, a full 68040 CPU is required to replace the default 68LC040 CPU.

* Classic family: SE30
* LC family: LC, LC II, LC III, LC III+, LC475, Performa 400-430, Performa 450, Performa 460, Performa 475, Performa 476
* Centris family: 610, 650
* Quadra family: 605, 610, 630, 650, 660AV, 700, 800, 840AV, 900, 950"

However, OpenBSD also classifies their entire Mac port as "Stagnant":

"The following platforms have been supported in the past, but releases are not being built for them at the moment. However, they are still living in the source tree, and releases might happen again in the future.

luna88k Omron LUNA-88K and LUNA-88K2 workstations
mac68k Motorola 680x0-based Apple Macintosh with MMU"

The Mac Linux page optimistically says the 660AV works:

http://mac.linux-m68k.org/status/MAC_MODEL_C660.php

The one thing you definitely cannot run is A/UX. The 660AV is also somewhat limited in what MacOS-es it'll run, being confined to 7.1-something to 8.1. (No System 6 or other such historical oddities.)

If you're looking to run weird OSes just for the sake of saying you've run them an old Pentium box would probably be a much better choice.

[Dog Cow]

sarcasm

Erm, only slightly. It would make a neat Retro Challenge project.

[ken27238]

It defiantly would

[luddite]

Something tells me an OS coded from scratch by one person in the space of a month might not be all that usable. I'm just saying...

[johnklos]

A 660AV is something of an odd duck and therefore used to be a bad candidate for running non-MacOS operating systems on. First thing that comes to mind is NetBSD, and NetBSD's machine status document:

http://www.macbsd.com/macbsd/macbsd-doc ... ne-status/

Says both ethernet and sound are broken on that machine.

I should really update that. Sound doesn't work on any of the m68k Macs in NetBSD except for little "blips" (for lack of a better word - kind of like a really short beep). Ethernet, on the other hand, has worked for ages.

The Mac Linux page optimistically says the 660AV works:

http://mac.linux-m68k.org/status/MAC_MODEL_C660.php

Unfortunately, Linux is not an OS (LNOS). Debian, which is an OS, has lost official m68k support, but a few people are still working on it. It's unclear how hard it is to bootstrap an m68k machine. Here's the latest status:
http://wiki.debian.org/M68k/Porting

I've been meaning to try to get the latest Debian on an m68k Mac so I can see how I can help, but the picture is not very clear. If anyone runs GNU/Linux on m68k, please let me know!

[Gorgonops]

I should really update that. Sound doesn't work on any of the m68k Macs in NetBSD except for little "blips" (for lack of a better word - kind of like a really short beep). Ethernet, on the other hand, has worked for ages.

I guess it doesn't surprise me a page or two would be a little out of date. The NetBSD project is rather "family owned and operated", after all. Note I say that with complete fondness, not as a swipe. For the last couple weeks I've had my old Ultra 10 busily compiling packages for sparc64 just because, well, it can. It runs Firefox 3.5 built from source acceptably, which is impressive for a 1998-vintage machine using an "odd" CPU architecture.

Admittedly I have to admit that in my opinion it must take a bit of masochism to run a machine as slow as a 68k Mac. I turned to the 440Mhz Ultra 10 to renew my acquaintance with NetBSD after deciding the 170Mhz Turbosparc-equipped SparcStation 5 I rescued from a recycle bin (and thus set this off) was just too slow, and that's a racehorse compared to a 68k Mac.

[Paralel]

Something tells me an OS coded from scratch by one person in the space of a month might not be all that usable. I'm just saying...

If the person has some exceptional talent anything is really possible. You might be rather surprised.

[Dog Cow]

Something tells me an OS coded from scratch by one person in the space of a month might not be all that usable. I'm just saying...

Are you trying to snuff us out? Trying to stifle competition?

[johnklos]

Telling me I can't do something is usually more motivation than telling me I can...

[luddite]

Then let me categorically state that it is impossible for one person to write an entire operating system in one month!

[johnklos]

Why, you...! I'll show you!

[Dog Cow]

Then let me categorically state that it is impossible for one person to write an entire operating system in one month!

You're really asking for it, aren't you?

[luddite]

my cunning plan seems to be working...

Anyway, now that this topic's been completely derailed... assuming you had the time, inclination and skill to write an operating system for 68k machines, what would be the key features that set it apart from what's currently available?

[ken27238]

lets say someone was going to write an OS. what language would be the most practical?

[Gorgonops]

C is about the most common choice these days but you could in theory use anything. Heck, write your own language while you're at it. Just remember that you're inevitably going to have to work your way all the way down to machine language/assembly for the really-low level things. (Like bootstrapping.)

Whatever the case, a 68k Macintosh would be a rather bad hardware platform to choose as your first every operating system project. Apple tended to use arcane hardware designs and poorly document them, since they never anticipated anyone really caring to access it outside of MacOS. It's notable that last I checked no one had actually written a "cold metal" bootstrap for BSD or Linux on the Macintosh. Booting them requires MacOS to initialize the hardware before passing control to the kernel via a bootloader that's actually a Mac program. Heck, even A/UX boots that way.

Of course, I seriously doubt this challenge is *serious*....

[johnklos]

my cunning plan seems to be working...

Anyway, now that this topic's been completely derailed... assuming you had the time, inclination and skill to write an operating system for 68k machines, what would be the key features that set it apart from what's currently available?

When I was a child, my third computer after a Sinclair ZX80, then Timex/Sinclair 1000 was a Sinclair QL. It came with 128k of memory, 512x256 four color graphics, preemptive multitasking, and an m68008. I was in love. This was in 1985 or so. Programming on the Z80 is cumbersome because of the fact that you're always painfully aware that you're on an 8 bit processor with some 16 bit abilities. I tried doing some 8086 assembly, but in reality it was no different than a Z80 with bank switching built-in. It was a kludge.

The m68k, though, was gorgeous. Eight 32 bit general purpose registers and seven 32 bit address registers (the eighth is the stack pointer) were a dream come true.

These days, kids are spoiled. They have tons of development tools for any kind of hardware they might find in the trash for free. You want a good c compiler? Download it. You don't have Internet? Go to the public library and get it. In 1985, though, compilers cost lots and lots of money. The only good c compiler for the Sinclair at that time was Small-C, which ironically was 16 bit (!) and cost a lot of money for a thirteen-year-old.

I ended up extending SuperBASIC on the Sinclair (it was actually not bad - it was typeless and procedural, and you could easily write code which looked very similar to the same code written in c or Pascal). I wrote programs with large parts in assembly language which took color screenshots and printed them on an ImageWriter II, extended that so I could print large color bitmaps with more colors (especially useful when I had a Trump card which extended memory to 896k), I wrote my own VT52 / VT100 terminal program with screen capture, redraw, printouts, and so on.

The nice thing about the Sinclair is that you could jump into assembly so easily from SuperBasic that it was completely easy to test, debug, and get things working in a short amount of time. Doing assembly, or even programming at all on a Mac requires (usually expensive) software and a bunch of setup to even get started.

If I were to write an OS from scratch, I suppose I'd like to use the model of the QL - a command line, a system with which to program easily, and an extensible procedural language so that one could start with primitives and build complexity in little pieces. It's very similar to the Unix philosophy, too.

[luddite]

If I were to write an OS from scratch, I suppose I'd like to use the model of the QL - a command line, a system with which to program easily, and an extensible procedural language so that one could start with primitives and build complexity in little pieces. It's very similar to the Unix philosophy, too.

Sounds good to me!

[Bunsen]

Then let me categorically state that it is impossible for one person to write an entire operating system in one month!

Oh I dunno, QDOS seems to have done alright for itself

[Anonymous Freak]

lets say someone was going to write an OS. what language would be the most practical?

English.

Followed by Chinese.

[luddite]

Oh I dunno, QDOS seems to have done alright for itself

It did even better for Microsoft!

[zerohour]

I am a sad man who actually planned an OS called CrossPlatform. The idea was a complete VP type OS so you could have Windows, Mac, Amiga and many other platforms on one desktop. Basically I know you could do it with emulators but the idea was to do the multiple pass interpretation like the DEC Alpha and make them all native so you could run them side by side with no performance loss.

Shame I can't code for toffee. It's also not practical but the thought was there. Running something like Wordsworth 6.0 on the same desktop as Maestro from the ST or such just appealed to me. Then again I may just be strange

[barana]

johnklos:

Why dont you try one of these? http://en.wikipedia.org/wiki/Q40