Submitted by The developers of RISC OS 5 are prepared to open source ‘elements’ of the operating system, acccording to a report in Archive magazine today. Castle boss Jack Lillingston reportedly told Archive editor Paul Beverley on-the-record that the company is “very seriously considering making RISC OS open source”.
…call it OpenDarwin – oh no wait, that’s taken.
(A little sarcasm – I assume 100% open source junkies are going to flame this “consideration” because it would not be 100% of the OS source code. I do not have a problem with opening only SOME of the code – its their prerogative).
Edited 2006-08-15 14:39
Yeah, it is sad to see that for open source junkies, the world is only black or white. Its so manichaean …
It’s sad to see that for closed source junkies, their stereotypes of open source/free software proponents are so black and well, black. It’s so boring…
I use both closed and open source, I just don’t care.
so you’ll gladly use the fruits of someone else’s labour (essentially their free gift to you), but have no hesitation in dissing them in the next breath.
if you really do like open source, (and if you use the internet at all and have an ounce of a clue how the thing was/is built, I think you would…), then you should appreciate that without some of that same “zealotry” as motivation you’re quick to condemn, little to none of this would have ever been done.
I don’t condemn open source, I condemn manichaeism (very popular in the linux community).
Why FreeBSD people doesn’t care if Apple release the entire system as opensource but not Linux people?
Why FreeBSD people are already porting technologies from OpenSolaris but not Linux people ?
etc…
Because they are not manicheans.
I give money to open source project I care about (OpenBSD) but I am not sure that the half of the linux zealots have given something to some open source projects (I don’t count their stupid flaming posts for something) …
While I don’t really care one way or another if an operating system is open source or closed, I do think if they’re going to open source it, with the hope of fostering outside development and increased community interaction then they should probably open source it all, be as open as they can to try and pull in the most possible interested people.
Really, I’d wish they’d BSD licence it, so it is used everywhere it can be used – though that’s not going to happen, one of the GPL-like licences will be the one used if they do it.
Really, I’d wish they’d BSD licence it, so it is used everywhere it can be used – though that’s not going to happen, one of the GPL-like licences will be the one used if they do it.
They can’t GPL-license it, since GPL means any piece of software linked to the GPLed software needs to be GPLed as well (including the parts they don’t want to open).
They might go for LGPL though; it would mean their competitor RISCOS Ltd, if they take the code, need to contribute back.
Either way, the idea that other OSes might benefit from the code seems farfetched either way. AFAIK RISC OS is filled with ARM assembly for optimisation, and if they open up just parts, those parts will probably be closely linked to other parts of the OS.
The restrictions of the GPL do not apply to the people who own the copyright to the software, only people using it under the licence – so they could GPL it, they’d just need to have a policy of submitted code being released to them, which isn’t uncommon with companies working with open source software.
…. or does it contain large chunks of ARM assembly?
Would be nice to see how it performs on PPC and x86 architectures.
RISC OS is an ARM-based OS. Just about all of the OS facilities are provided through SWI calls (SoftWare Interrupts) which are an ARM mechanism.
The way that SWIs work is basically the bottom chunk of memory is a jump table to certain OS routines. IIRC when the processor comes across a SWI instruction the it (assuming the conditions for the instructions are met) changes its execution mode, looks up the relevant address for the SWI from the table, and jumps there.
The SWI system is extensible – you can load new OS modules in that provide more SWI calls.
RISC OS is basically divided up into a load of different modules which implement SWIs. I expect that Castle are thinking of open sourcing a chunk of the OS modules that make up RISC OS 5. There are probably a number of modules that they cannot open source due to legal reasons.
Disclaimer: It’s been about 15 years since I wrote a RISC OS application, so I may be a bit rusty on some of these details.
Sounds like it takes advantage of hardware dependent features on ARM. But then, Linux and NetBSD take advantage of hardware dependent features on the platforms they run on, too – even if constrained to provide a UNIX interface (so privilege separation, please, and no segmentation, thankyou).
Most processors have a software interrupt mechanism and using a software interrupt is the generally accepted method for a user land application to make a call into the kernel.
I haven’t touched one in years either, but I think the basic ideology is still there.
The current state of RISCOS is terminal as Castle and RISCOS Ltd won’t talk with each other and they cannot get enough interest to get an order for new hardware (a few months ago it was big news that some old motherboards were found in a wherehouse somewhere, that’s how bad it is!)
Open sourcing at least some of the OS makes sence as both parties are struggling to update some of the modules to work on the new OS’s, so why not open some of the less important/most tricky/most complained about bits and let the community do the work (and they probabaly will).
If this is a hit then it would not suprise me to see about 90% being open after a while (leaving the BIOS, kernel and a few other bits closed to stop the users forking to a totally independant version).
So a port to another architecture would make it RISCOS only in name, if I understand correctly?
nah, what makes RISCOS RISCOS is really the use of middle click popup menus for app menus (File, Edit), its modularity, and probably most important its very ubiquitus use of drag and drop
even if its infeasable to port to different processors theres still a wealth of ARM products out there it can’t run on at the moment that it could be ported to.
getting riscos on a PDA would be cool especially the massive use DnD seems like a good fit for a stylus based interaction
Edited 2006-08-15 17:54
If they open source the odd bit here or there, then it won’t make a ha’p’th of difference; but (assuming that Castle are losing interest), open sourcing RISC OS could make the difference between seeing it die and seeing it pick up a user/developer or two.
I’m of the belief that open-sourcing ANYTHING is signficant, because it promotes the notion that open source is the ultimate destiny of almost ALL useful software.
Significant for open-source generally, yes. However, as the article points out, just because something is open source, doesn’t mean it’s an immediate and guaranteed winner. It’s just not guaranteed to die with the company that made it.
> open source is the ultimate destiny of almost ALL useful software.
Gee!! While I am an open source enthusiast and developer myself, I wouldn’t wanna live in a world like that
One Paradigm to Rule Them All… it sounds too MicroSoftish to me.
I, like many other open source enthusiasts, like choice. And I like the choice of having my software being developed either in a proprietary or free license.
Every situation is unique guys.
Bear in mind that RISC OS is a highly modular operating system. If some of the modules were open sourced it would allow people to update them and be plugged straight back into the system to benefit everyone. Hence even if some of it were open sourced, it could still be of great benefit to users.
This is true but I am more excited at the prospect of enough of the OS being opened up to make a port to another platform possible. PDA RISC OS here we come!
If its open source, and easy to get hold of. Plus its not tied to a specific platform. Then I’d be interested, otherwise it’s another morphos.
My last experience was playing that David Braben demo! Man that was years ago!
If it is just one more dying OS that the company really can’t make a dime on I really truly belive that the best thing to do is to open source the whole thing, otherwise it will just end up dead and forgotten:(
Imagine if Commodore had open sourced AmigaOS just before it died, then it might have been a true contender today.
I agree that one of the important consequence of open-sourcing RISC OS would be that it would be impossible for it to disappear overnight.
However, hopefully Castle are nonetheless still making money from it. From their point of view, there has been some indication that making it open source would further help them promote it as an embedded OS (OEMs are apparently happier using open source software as it gives them more control; see http://www.drobe.co.uk/riscos/artifact1671.html ).
Similar facilities to the ARM SWI mechanism should be available on any modern hardware platform. However, if (as seems likely) there is inline ARM assembler implementing these, some code refactoring could be in order to make the OS runnable on other hardware platforms.
that modern XScale systems are a bit overpriced..