RISC OS is being relicensed under the Apache license | Hacker News Reader

RISC OS is being relicensed under the Apache license

riscosopen.org

98 pointsmesse41 comments

When I first came across RISC OS years ago, I couldn't help but be amazed that the entire OS ran in only 4MB - and often burned to ROM[0]. Just the write-protection alone made this impressive to me - especially as something useful to the general population.

A short while later, I came across an effort to run a riscOS-based computer (circa 2001), basically on batteries[1]. It seemed to solidy the 'light on resources' aspect of the entire riscOS stack.

RISC OS usage[2] seems to indicate it's a little tricky (non-intuitive) for first-time users[2]. Hopefully if someone picks up development on this, improving the interface a bit (ie, 3-button mouse reliance, etc), plus adding wifi are things that are tackled first.

Honestly... would hate to see feature-bloat creep in too much. I played around with KolibriOS[3] the other day and there's certainly a benefit to small and fast!

[0] https://en.wikipedia.org/wiki/RISC_OS

[1] http://www.explan.co.uk/hardware/solo.shtml

[2] https://www.makeuseof.com/tag/turn-raspberry-pi-retro-pc-ris...

[3] https://www.kolibrios.org/en/

RISC OS was the majority OS in British schools for part of the 1990s. It's not difficult to use, though it does have some different paradigms. Or rather, it went 100% for some UI decisions which other OSs only have as secondary features.

Middle click opens a menu (like a right click menu on Windows). Usually this is the main menu, the context will change depending where you click -- to get a save option, you need to click on an open document.

Everything is accessed through the file browsing windows. Opening document and applications, and saving documents (dragging them into an open window). Drag and drop is very common.

I think that's all :-)

Nice.

So RISC OS is now open source. EmuTOS is an open sourced Atari ST TOS. Hell, Apple is open sourcing LisaOS.

Why isn't AmigaOS open source yet?

I get the feeling the AmigaOS people still think they can sell it or they have a bunch of code that they know will be problematic to open source.

Either that, or they don't have some of the code anymore. It's changed hands a few times so it's possible that some of it is only available as binary files packed into the ROMs and other places.

Hyperion released a new version of AmigaOS just a couple of weeks ago http://hyperion-entertainment.biz/index.php/where-to-buy/dir...

AmigaOS 3.x source was leaked on github a few years back. Quite a few people have copies, though as you note is likely not 100% complete.

One of the devs (as opposed to owners) is busy with DragonflyBSD.

"Matthew Dillon, an Amiga developer in the late 1980s and early 1990s and FreeBSD developer between 1994 and 2003, began working on DragonFly BSD in June 2003 and announced it on the FreeBSD mailing lists on 16 July 2003." - Wikipedia, DragonflyBSD

I thought he had developed the DICE C compiler, not parts of the OS.

"Dillon started DragonFly in the belief that the techniques adopted for threading and symmetric multiprocessing in FreeBSD 5[4] would lead to poor performance and maintenance problems."

So whether he contributed to the OS or just learned something at Amiga from those who did, Amiga's key concerns live on. Not to mention that the compiler isn't exactly remote from multiprocessor concerns; its value was that it adapted closely to what the Amiga hardware made possible, obviously.

”Apple is open sourcing LisaOS”

They will allow the computer history museum to host the source code, but chances are it won’t be under an open source license. That’s what they did before. QuickDraw and MacPaint have a ”available only for non-commercial use” clause, for example (http://www.computerhistory.org/atchm/macpaint-and-quickdraw-...)

And AmigaOS? Does anybody know who owns the rights to that? ;-)

Yeah fair enough re: LisaOS. I also don't know what's happening to it, since it's been almost a year since the announcement and not a peep about the code. I was looking forward to reading through the code.

And Microsoft released MS-DOS 1.25 and 2.0 under the MIT license.

> Why isn't AmigaOS open source yet?

Possibly because someone is still exploiting derived code commercially?

Anyway the 1.2/1.3 codebase is in large parts BCPL, not C.

Current rights owners of Kickstart and Workbench (I don't even know what company owns that now, it's being reselled again and again) are aggressively trying to extract last money from it.

Also, there's hatred to open source in general in Amiga scene. Interesting comment in one discussion about MorphOS: https://lobste.rs/s/vkpekg/morphos_3_10_released#c_ktgnn3

"the mission to reinvigorate the RISC OS market"

They're nearly 20 years too late for that. The only compelling reason to use RISC OS over anything else these days is nostalgia. Still, it might be an interesting project to hack on.

Running RPCEmu, I was surprised how real the nostalgia for old operating systems can feel!

( and I still haven't gotten Digital Symphony to work on any kind of modern machine. :-/ )

I converted a lot of my music from SoundTrackers to Digital Symphony files because it had compression, so I needed half as many floppies. Now, with nothing to play them on, I regret that.

That's terrible to hear! Perhaps you'll find some of them back here: https://modarchive.org/index.php?faq-torrents

Also, xmp will play symphony files, http://xmp.sourceforge.net/

(installs fine on debian with apt install xmp )

If you just want to listen to your mod collection again, that might be sufficient already! :-)

I had a look at the Icon Bar site [1], and I'm amazed that people are still able to sell software for RISC OS.

I was given an informal (but firm) takedown notice in about 2002, when I uploaded an old game to my website and linked to it from a RISC OS forum. I'd thought RISC OS was dead then.

[1] https://www.iconbar.com/

I don't know about that. There's a serious market for a real desktop personal computing operating system right now, one that is simple (note: not easy, simple), and enables users instead of trying to coddle them. Could RISC OS fill that role? Probably not without a whole lot of changes, starting with a conversion to preemptive multitasking. And worse, if anyone took it up they'd probably make it all POSIX and integrate a package manager and try to turn it into another Linux Distro.

Still, I can hold on to vain hopes.

> There's a serious market…

The personal computing market has already been divided between Windows, MacOS, and ChromeOS.

I think simple and useful are competing goals. RISC OS is simple, but there's a lot that it can't do. I expect that simplicity makes porting larger software projects pretty difficult too.

> starting with a conversion to preemptive multitasking

I wonder how difficult that would actually be. I think it could be done without changing the existing WIMP API. The Wimp_Poll loop would still exist as is, just as in other windowing systems; it just wouldn't be the only time it loses control.

And preemption has certainly been done before in Niall Douglas's Tornado project.

> I think simple and useful are competing goals.

Counterpoint: the stick, the knife, and the wheel.

How often do you use a single stick or wheel, that isn't part of some much more complicated machine?

(I won't argue about the knife. That is both simple and useful)

That doesn't seem like a counterpoint at all. For example, a knife would be more useful if you could also eat soup with it, but it would be less useful.

Why would I want to eat soup with a knife? That sounds like it would cause more problems than it solves. A knife you can eat soup with is exactly the kind of over-engineered nonsense feature that modern OSs cram in. But I could carve a spoon with a knife.

Let me give a less analogous example: Windows recently (or will soon, I forget) have a Dark Mode, which will make window colors darker and text colors brighter. This is a feature upgrade, apparently. Yet back in Windows 95 people would have laughed at you for making a big deal out of that, because you could just arbitrarily change all the window and text colors at will. No reghacks, no CSS, just a simple settings widget [0].

Another example is how ludicrous application installation has become. In the DOS and MacOS Classic days you just copied a folder to a disk, any disk, and ran it and it worked. Nowadays there's package managers and SxS and you can't even install applications to different disks. Is the former method more limited? In some ways yes. But it is also much more flexible in others and much simpler, it's easy to reason about and fits perfectly with the file management metaphor that is the core abstraction of desktop OSs.

[0]http://toastytech.com/guis/win95schemes.png

I wish that the QNX OS was open source and free as in beer.

How far away is this from being as usable as I hear QNX is?

Usable for what purpose? They're about as dissimilar as could be.

I don't think they have much in common. There's not much out there really similar to QNX. Maybe Minix, since it's a microkernel with good userspace POSIX support. No "real time" support though. There's FreeRTOS, but it's not terribly similar to QNX either.

Thanks! My brain seemed to think the R in Reduced Instruction Set was for real time ...whoops!

Neat project regardless.

Interesting choice of license. It will be very interesting to see what kind of growth their community experiences as a result.

ARM only, single-user, cooperative multitasking, 31 bit max file size and somewhat limited memory protection. I suspect the audience would be limited to niches where all of that isn't an issue.

Building an rpi-based education platform might be a good candidate. Though I'm not sure what the killer feature there would be, and there's already the pi-top in that space.

It might appeal to those people who are nostalgic about the sort of elegance and simplicity of the design that went into MS-DOS/Windows 3.1. RISC OS is not quite like that, but it is a fantastic system in its own right, especially since it runs natively and fast on the Raspberry Pi. I hope that this move will create a huge surge in community support of this beautiful OS.

I wonder what the vision is for what they want it to become?

Cause realistically most of the OS bases are covered now.

This should allow "GPLv2 or later" code to be used, right? (Allowing it to be ported to more ARM platforms)

I don't believe so. Here's the Apache project's position on the subject:

"Apache 2 software can therefore be included in GPLv3 projects. . . However, GPLv3 software cannot be included in Apache projects. The licenses are incompatible in one direction only"

(https://www.apache.org/licenses/GPL-compatibility.html)

(edit: Forgot to mention GPLv2 - that one is even less compatible: "Despite our best efforts, the FSF has never considered the Apache License to be compatible with GPL version 2")

Yea, I know it might require the RISC OS kernel to be relicensed.

If you are talking about the Linux kernel. Then yes, a kernel relicense is absolutely required.

I have received a few downvotes. Please note I was answering the above question that a relicense is absolutely required for compatibility with the Apache v2 license because of the patent provision that is incompatible with the GPL v2. This is the position of the Free Software Foundation. The GPL v3 has a patent provision and explicit compatibility with the Apache v2 license to end any ambiguities.

Reply on news.ycombinator.com