RISC OS 5 source code release revealed | RISC OS news and information from Drobe


Beta! \| About us \| Contact \| Submit news \| RSS \| Twitter	Webspace \| Tech docs \| Downloads \| BBC Micro \| Gallery \| Wallpaper

RISC OS 5 source code release revealed

Published: 29th Sep 2006, 21:55:39.

[Updated] Reaction and views from across the platform

Castle and RISC OS Open Ltd have revealed their ambitious plan to release the source code of various RISC OS 5 components. Under a so-called 'shared source initiative', a phased release of RISC OS is promised to take place over the next few months - starting with applications including Paint, Edit, Draw, Configure, Unicode, the web browser Browse, and the printer manager. The Shared C Library and RISC OS build scripts will also be released, according to CTL and ROS Open.

It is hoped the project will kick start outside development of the desktop operating system, which has somewhat died down over recent years. It is also hoped the various hurdles that minority closed source operating systems face - such as modern driver development - will be more easily tackled with extra developers working on the source code.

But independent programmers have so far viewed the announcement with mixed opinions. While some say it is hard to judge the impact of this new initiative without seeing a formal license, others have warmly welcomed the news and pledged to take up development.

Main points
- The RISC OS source code - the blueprints to its design - will be made available, and can be updated and re-released by third party developers.

- The source code will still belong to Castle. A formal 'shared source' license will be published "in due course".

- Applications can include RISC OS modules provided developers publish the modules' source code.

- RISC OS Open hope to "promote more rapid development of the software base", and say support and encouragement will speed up the roll out of code.

- They will set up a system to accept updates and changes from coders.

- As exclusively revealed earlier this year on drobe.co.uk, RISC OS Open are staffed by ex-Acorn, ex-Pace Micro, ex-Element-14, and ex-Tematic engineers and managers.

- Over a million deployed consumer electronics products are RISC OS powered according to Castle and RISC OS Open.

RISC OS Open have stressed that the license will not be a traditional open source agreement. Instead a dual-license will be employed allowing people to view the software blueprints, make modifications, and distribute updated components provided the updated source code is handed out too. To use the RISC OS source code commercially, you must pay a royalty fee.

There is no requirement to pass source code changes back to RISC OS Open, and while the company will act as gatekeepers to the official source code, this will be performed on a not-for-profit basis. Punters will be able to donate cash to the project.

In a statement, RISC OS Open company secretary Steve Revill said: "It's the first time that the public have been able to access these components at a source level. We believe that it's an exciting opportunity for developers to submit their improvements for all to use.

"With the help and support of the RISC OS community, I am sure that there will be even more to follow. It is our sincere hope that this will stimulate the growth of our community and introduce a new era of development."

Castle boss Jack Lillingston said: "In today's era of fast moving technological developments, both users and developers should have worldwide and easy access to an operating system. For too long RISC OS has been regarded as a closed proprietary OS and this has hampered wide scale take-up.

"We are delighted to be supporting the creation of an international RISC OS community with the team at ROOL. Their activities will further the use of RISC OS in a structured way, promoting the take-up of RISC OS."

Reactions
RISCOS Ltd boss Paul Middleton said: "We are too busy finishing off the first Select 4 issue to comment much on the matter.

"It is probably worth pointing out that the 'open sourcing' of RISC OS is going to solely cover RISC OS 5 versions. We do not intend to 'open source' RISC OS 4 versions as some people seem to have assumed.

"I would point out though that we have always been happy to work with developers who require source level access to RISC OS, in the same way that Acorn made sources available for particular projects. The difference between us and ROOL is that we do require any changes made to be fed back to us, as we only want one version of RISC OS 4 to be available."

Aemulor and Geminus developer Adrian Lees said: "I'd want to know if there's no 'contamination' clause so that I could still go ahead and re-implement an OS component later after seeing source code, if I so chose. Of course, I want to know what will be open sourced; I'm especially interested in lower-level software - the Kernel and Window Manager.

"It'd be nice if this and new hardware development from Castle fitted together. If CTL does make a new motherboard, I'd happily contribute to the OS changes needed to get it running.

"Also, supposing I plus someone else wanted to port RISC OS to different hardware, bought in or custom, I want to know what the license fee per unit is likely to be, ie. how much we'd have to add to the hardware price as a minimum to cover costs. is there a clause about the target platforms, especially emulation on non-ARM CPUs?"

CDVDBurn author Steffen Huber said: "I really welcome this move. To me, it is the first positive piece of news for our market since the launch of the Iyonix. It is probably the only way to keep RISC OS alive, and it has the potential to interest many ex-RISC OS developers.

"It is of course too early to consider all consequences - what exactly is that 'per unit royalty fee', would this 'shared licence' allow royalty free shipping of a freeware emulator complete with RISC OS, does selling support for an otherwise free product make it a commercial product. The coming discussions will keep us surely busy for a while."

Artworks and Techwriter developer Martin Wuerthner said: "I am delighted to see progress in this direction. I was about to ask Castle for the source code of some OS components because I wanted to release improved versions, so this move is very welcome."

GCCSDK developer and Firefox porter Peter Naulls said: "Some of the choices are a bit strange - Browse for example. Do we really need to resurrect that, with respect to the considerable effort that Acorn developers did put into it? On the other hand, the core apps - Paint et al - and I suspect I won't be the last to say this, have more than a passing resemblance to those which had considerable improvements by RISCOS Ltd for Select and Adjust. This might be another nail in the ROL coffin.

"I remain cautious about this, until I see precisely what's available and of course what the precise conditions of the licenses are, and bearing in mind that much of RISC OS continues to need is application development, not OS-level improvements per se, and the distraction that this might bring. On the other hand, it will no doubt revitalise some developers."

R-Comp boss Andrew Rawnsley was more reserved. He said: "It seems fairly reminiscent of RISCOS Ltd open sourcing parts of Printers, ie. not hugely exciting. I guess if people are interested in developing those components, it is good, but really, you have to ask, shouldn't Castle actually have paid coders working on these things, especially given the open-but-not licence?

"Unfortunately, most of the items listed have better versions already in existence as part of Adjust, so really all it does is accentuate the OS split, because the public only have source access to 'out-dated' versions of things."

Rob Kendrick, NetSurf GTK developer and drobe.co.uk contributor, said: "I assume CTL are hoping or praying that suddenly thousands of companies will start using RISC OS and give them royalties. From what they're suggesting, it really is Open Source, just with a no-commercial-use clause in one of the two licences the source is available under.

"The only people who use 'shared source' are people who hate 'open source' - ie, Microsoft. The two are essentially the same."

NetSurf developer John-Mark Bell said he would reserve judgement until he has seen the exact 'shared source' license.

Michael Drake, NetSurf designer and contributor, said: "This sounds good. My view is that anything like this is better than nothing. RISC OS was going nowhere. And the RISC OS Open website is a RISC OS site that doesn't look vile."

Currently, no fixed release timetable is known, and the exact license conditions have yet to be published. RISC OS Open first popped onto the radar in June, prompting questions of what would opening up the OS achieve - Castle later dropped hints at its plans to release the blueprints. RISCOS Ltd's Paul Middleton argued earlier this month that Castle's wish to reveal the OS source code would have no effect on ROL, and that this "is not the panacea for suddenly taking RISC OS forward. No operating system can be developed solely for free."

• Earlier today Jack declined to comment further on how Castle may use the new IOP XScale cores. He said: "You know us - no comments about future products."

Links
RISC OS Open website - FAQs, the lot
Announcement PDF - drawn up in Microsoft Word

Discussion

Viewing threaded comments | View comments unthreaded, listed by date | Skip to the end

Been checking the RISC OS Open site, but not seen any changes? Clear your browser cache! Reload:

[link]

:-)

adh1003 on 29/9/06 9:58PM [ Reply | Permalink | Report ]

Perhaps ROOL might be get a theatre slot or stand at the SE show as I am sure there will be lots of questions....

markee174 on 29/9/06 10:13PM [ Reply | Permalink | Report ]

Hopefully this will kick start a few things and result in a few updates for use end users.

SE Show should be intresting just for this, even if ROOL are not there I'm sure castle will be able to talk about it. Also what does this mean for ROL's Select for iyonix as to me this cuts yet another excuse away as they can now have access to the source code to make the kernal changes?

bluenose on 29/9/06 10:26PM [ Reply | Permalink | Report ]

This will be my first drobe comment ever; I got really excited by what must be the best news for a while : fast new hardware and open sourced OS. Excellent!

fantasian on 29/9/06 10:28PM [ Reply | Permalink | Report ]

So now we are going to have RO5, ROAdjust and ROxx with various bits of the OS open sourced. How much sense does this make?

sa110 on 29/9/06 10:29PM [ Reply | Permalink | Report ]

okay, re-reading the article, it would appear my earlier comment may have been a bit rushed. Allowing the devlopment of various OS based applications and Browse is a good thing.

However when it gets to updating the OS itself, I standby my earlier comment.

sa110 on 29/9/06 10:34PM [ Reply | Permalink | Report ]

I ought to just reiterate that this isn't technically open source, although the source code is out in the open for the public to use. The strict definition of open source doesn't allow dual licences, whereas what we're doing at ROOL does - commercial users have to pay a royalty. That's why we're using the term "shared source" on the web site.

adh1003 on 29/9/06 10:37PM [ Reply | Permalink | Report ]

In Reply to sa110

Well then again perhaps we could have on joint RISC OS if someone merged the ROAdjust and RO5 that run on both the Iyonix & A9 not to mention any new hardware. Who knows ROL might do that work. Still I'm getting a little carried away and at the moment only some sources are being shared sourced but as "fantasian" said this is an exciting development .

bluenose on 29/9/06 10:38PM [ Reply | Permalink | Report ]

This is good news. I look forward to seeing the sources to be made available and to experimenting with them. I also look forward to seeing what others will do with them.

GavinWraith on 29/9/06 11:16PM [ Reply | Permalink | Report ]

adh1003 wrote: "The strict definition of open source doesn't allow dual licences..."

Where is the strict definition of open source that says that?

davidb on 29/9/06 11:39PM [ Reply | Permalink | Report ]

adh1003: The point about this not being 'Open Source' is made on ROOL's website. I was rather argumentative in a previous discussion about a proposed license, having objected to it being called Open Source. I'm glad that ROOL have made it nice and obvious on their website.

flibble on 29/9/06 11:53PM [ Reply | Permalink | Report ]

Look at the OSI web site [link] - there is a link to a definition of Open Source over on the left. You should also read the FAQ on the [link] site.

7thsoftware on 29/9/06 11:54PM [ Reply | Permalink | Report ]

I guess the strict definition of open source is about, well, open source. However, the dual licensing thing just adds another dimension - it doesn't take an existing dimension away. That something is available under a closed source licence doesn't mean that the same thing, licensed under an open source licence, is suddenly not open source for all those people who have received the software under the terms of such a licence. Indeed, commercial agreements often run in parallel to the availability of software as open source and do not compromise that in any way, and I'd be surprised to see any open source definition expanded to cover what is effectively a business strategy.

Anyway, the "shared source" licensing of RISC OS "Open" seems like an odd combination of GPL-like distribution terms with a commercial exploitation restriction. It'll be interesting to see what effect that restriction has, and whether the GPL won't ultimately prove to be a better choice.

guestx on 30/9/06 12:35AM [ Reply | Permalink | Report ]

This news has really made my day. The licence may not be truly open source, and all of the details may not be available yet, but what does seem clear is that the source code for many of the RISC OS components will be available for everyone to access for the first time.

It'll be great to have the opportunity to look at and change the underlying functionality of the OS. This opens up a host of possibilities

You look at many of the suggestions people made under the Merlin scheme or have made on the Iyonix list, and whereas before people had to speculate as to how things might work, it will hopefully now be possible for people just to go out and try things.

flypig on 30/9/06 2:32AM [ Reply | Permalink | Report ]

> Anyway, the "shared source" licensing of RISC OS "Open" seems like an odd combination of GPL-like distribution terms with a commercial exploitation restriction. It'll be interesting to see what effect that restriction has, and whether the GPL won't ultimately prove to be a better choice.

I'd argue that since GPL in /practice/ if not in letter prevents the code being used in commercial products, that the proposed dual-licensing arrangement (at least as currently outlined) is actually more flexible.

adrianl on 30/9/06 3:28AM [ Reply | Permalink | Report ]

adrianl:

Why not release it under GPL and also an alterative license as MySQL does?

markee174 on 30/9/06 8:31AM [ Reply | Permalink | Report ]

Would this not mean ROL could release the same components (as they tried with !Printers) and the bug fixes/improvements could be itegrated into a combined source ?

tank on 30/9/06 8:37AM [ Reply | Permalink | Report ]

markee174: "Why not release it under GPL and also an alternative license as MySQL does?"

Because obviously they don't think they can sell services & support, or provide enough value-add on top of a truly open source product.

And in the commercial market that Castle's aiming at (embedded devices) the GPL - and other open source licences - have been very succesful. In those markets, the OS without the hardware is generally useless; but many commercial, embedded, hardware products use GPL software with no disadvantages (and sometimes advantages). Think of all those routers running Busybox, a Linux kernel etc. etc.

*Personally*, I'd've thought that an STB maker would want support for their OS, and so value-add services on top of an open source product would be viable. But presumably only if the customer trusts the supplier to be able to supply those services in a timely and efficient manner.

Anyway, I'll stick my 2p in too: it'll be interesting to see how much of a running system can be replaced. There's no mention of the kernel, or most of the core of the OS; so I'd imagine we're most likely to see developers picking up things like !Paint to improve them.

Jaffa on 30/9/06 8:39AM [ Reply | Permalink | Report ]

Jaffa:

Haven't you just argued in favour of releasing the code under both a GPL and a non-GPL license?

The GPL version encourages people to look at it and investigate the source code. The serious customers will want the non-GPL version and support - the GPL version just provides a better way to reach them.

markee174 on 30/9/06 8:44AM [ Reply | Permalink | Report ]

What about the improvements already made but are part of RO4 Select/Adjust ? Unless the same modules and apps are open sourced from there as well then somebody will simply duplicate the work that has already been done.

I'm all for open sourcing but I see some serious pitfalls due to the split OS fork. I am also a bit suspicious that this is an attempt by Castle to drop development itself for those modules and apps that it knows are inferior to the ones in Select. Has anyone checked this ?

mripley on 30/9/06 12:17PM [ Reply | Permalink | Report ]

This is the best news we have had on the future of our moribumd OS for a very long time and prompts this, my first comment to Drobe. I am greatly encouraged by the positive reactions of Developers and others whose achievements and abilities I greatly respect, even though there is a lot of water to flow under the bridge before we know what the outcomes will really be.

bric2 on 30/9/06 12:38PM [ Reply | Permalink | Report ]

Yes, this is what I am worried about. Given that alot of effort has been put into improving Paint etc, as part of Select/Adjust it seems a waste for this to be duplicated.

My other concern is about what people can actually get and what people will still need to pay for. Will it be possible to get hold of the full OS, or would you still need to buy this? I can't see people being that keen to develop the OS if it is the latter.

Walks on 30/9/06 12:43PM [ Reply | Permalink | Report ]

Not that I want to p*** in the pool, especially on top of this latest shot of chlorine, but didn't Castle get a bit miffed a while back when RISCOS Ltd Open sourced Printers? Apart from that, it will be interesting to see what effect this has on RISC OS. I wonder what RISC OS Ltd think about it? Cheers!

fwibbler on 30/9/06 2:55PM [ Reply | Permalink | Report ]

markee174: "Haven't you just argued in favour of releasing the code under both a GPL and a non-GPL license?"

Not necessarily. Selling services around a wholly GPL product is perfectly possible. Dual licencing may make more sense if you want to value-add software, but you need copyright assignment.

"The GPL version encourages people to look at it and investigate the source code. The serious customers will want the non-GPL version and support - the GPL version just provides a better way to reach them."

Except the GPL doesn't allow you to distinguish based on usage, so commercial customers who don't want support can use it without paying a royalty.

Jaffa on 30/9/06 2:55PM [ Reply | Permalink | Report ]

adrianl: "I'd argue that since GPL in /practice/ if not in letter prevents the code being used in commercial products"

I guess all those Linux-based devices are purely imaginary, then.

guestx on 30/9/06 3:04PM [ Reply | Permalink | Report ]

This could be very good news for RISC OS in general. There are certainly short term opportunities this will bring to ensure new APIs are available on all platforms, and I hope that over the long term with the continuing enthusiasm (or is it now masochism

) of RISC OS developers, to produce significant new functionality for the OS.

For example in the short term the unicode font manager can be made available for all platforms, bring the benefits of NetSurfs support for foreign character sets to all, and making life easier for its developers. If the print manager also contains the drivers, it will be possible to patch the missing support for unicode text printing in RISC OS 5 at the same time making it available for 4.x.

One thing ROL doesn't appreciate is that by not making APIs universally available, developers will be reluctant to use them, and without programs either requiring people to upgrade, or at the last tempting people with the promise of additional functionality if they do, then there just isn't a good case for buying an OS upgrade. ROL now has the ideal opportunity to licence some already written code and get out a release to Select customers, however I suspect the reluctance they have shown in the past to paying licence fees will mean that miss this boat too.

However, the good news for users is that if ROL don't commercially exploit it, other developers can still make it available under the shared source licence, and any improvements can be returned to ROOL for the benefit of everyone.

druck on 30/9/06 3:32PM [ Reply | Permalink | Report ]

> I guess all those Linux-based devices are purely imaginary, then.

I was thinking of products that are purely software. It's a little more difficult to copy and distribute hardware.

adrianl on 30/9/06 3:35PM [ Reply | Permalink | Report ]

To Peter Naulls: "Some of the choices are a bit strange - Browse for example." Browse is a good idea since it is still my favourite Browser for offline use and even for quit a bit online use if I surf using RISC OS. Getting a 32 bit version would be nice indeed.

To druck: I like your comment about ROL :-)

So let us hope that this attracts the odd developer to add features to RISC OS 5 now ... or perhaps ROL now gets to know how to offer Select for the IYONIX ;-)

hzn on 30/9/06 4:05PM [ Reply | Permalink | Report ]

adrianl: "I was thinking of products that are purely software."

Well, it's possible to sell GPL-licensed code, of course, as well as to provide paid support and services for such code whether obtained for free or otherwise. And there are real people out there who seem to have made such business models work. Meanwhile...

druck: "ROL now has the ideal opportunity to licence some already written code and get out a release to Select customers, however I suspect the reluctance they have shown in the past to paying licence fees will mean that miss this boat too."

The problem with this scheme is that companies like RISC OS Ltd. will see that they're automatically lower in the hierarchy than Castle, and since they'll want to keep the end product closed source, they may well avoid developing anything that passes revenues upwards in that hierarchy. Even if they want to sell stuff whilst giving away the source, they still have to pay Castle something.

guestx on 30/9/06 5:24PM [ Reply | Permalink | Report ]

Browse is an excellent application to get working again (my copy crashes and effectively kills my machine). This is surely the browser that is closest to what we require of a modern browser. Netsurf is excellent however it does lack Javascript which Browse has.

Another advantage of developing Browse is that there is nothing "owned" by ROLltd to conflict with any development work. It would also plug one of the biggest holes in RISCOS's lack of appeal to the outside world.

mripley on 30/9/06 5:45PM [ Reply | Permalink | Report ]

Good positive news, to be welcomed.

This hopefully *will* encourage more development of the RISC OS 5 and it's capabilities (both in supporting new hardware, a new motherboard should CTL produce one, and also in UI elements where it might be argued Select has the edge).

To some extent RISC OS developers now become *stakeholders* in the OS - and that sense may well encourage more substantial developments of the OS and even the platform.

In my humble opinion *good news*.

AMS on 30/09/06 5:54PM [ Reply | Permalink | Report ]

Jaffa:

"Except the GPL doesn't allow you to distinguish based on usage, so commercial customers who don't want support can use it without paying a royalty."

Not directly. But it forces the GPL to be applied to all derived software which would limit its practical usage in many cases. So clients would use the commercial license.

The big question is how to reach potential commercial clients and by vastly increasing exposure whilst limiting usage, GPL can help with this.

markee174 on 30/09/06 6:04PM [ Reply | Permalink | Report ]

I can't see how this can be bad news, but I hope we're not kidding ourselves about how 'good' it is - ie the extent to which it will spark a new wave of development for the OS. If it generates more interest in commercial and cross-platform use that will be great, but I wonder if other more developed systems (WindowsCE, Linux etc) haven't already occupied a lot of that ground.

Incidentally, how do you get hold of Browse? I've never seen it available for purchase or download.

bstewart on 30/09/06 6:41PM [ Reply | Permalink | Report ]

Somebody should look around and find/offer Paul Middleton a new job. Then he could afford to make all the select code open source too. After all the developements have been paid for by users. I am sure the subbscribers would vote for that. I suspect that he will try and carry on, which just means a fork in the OS. Would it not better if all subscriptions went into a prize pot that could be handed out to worthwhile developments. Would at least take out the overhead of having to pay Paul Middleton as manager.

Jwoody on 30/09/06 7:46PM [ Reply | Permalink | Report ]

This shared source license looks at first glance very similar to the kind of thing that MySQL, KDE and so on do, I'm not sure why some people think it is different.

I had to stop dancing rount the room and kissing my RISC PC to write this

steelpillow on 30/09/06 8:13PM [ Reply | Permalink | Report ]

steelpillow: ROOL themselves freely admit it's very different to the open source licence KDE is under. The main, most obvious example being that any hardware manufacturer could put together a box with a completely open source software stack (from BIOS to office suite) and sell said hardware.

This isn't possible with ROOL's shared source licence (even assuming that eventually enough components are released to build a completely working system without any dependence on an existing underlying RISC OS)

Jaffa on 30/09/06 8:28PM [ Reply | Permalink | Report ]

IMHO, it will be entirely possible to build a box using open RISC OS components and sell it. You would have to pay a unit royalty to Castle, but that would be your one and only notable restriction.

adh1003 on 30/09/06 8:53PM [ Reply | Permalink | Report ]

Andrew: sorry, I should've been more explicit. It's possible to put together a completely open source software stack machine and sell it *without paying any royalties*.

Doing the same with ROOL's shared source would, as you say, only be possible when paying Castle a royalty.

However, it's interesting you say "entirely possible to build a box using open RISC OS components and sell it." Is it the intention, then, that ROOL will eventually share enough to create a bootable system entirely from source?

RISC OS 6, free for end-users, running on RiscPCs is an appealing prospect, given the RPC still makes up the majority of current hardware.

Jaffa on 30/09/06 10:00PM [ Reply | Permalink | Report ]

hzn: That comment was made on csa.programmer, and you've replied here, with no indication that it was so, nor where to look for what else I said if others cared about the context. If you care to reply there, I will be happy to respond.

mrchocky on 30/09/06 10:29PM [ Reply | Permalink | Report ]

dual licensed software, including KDE and MySQL tends to be GPL + something else. Even among professional software developers, there is a lot of misunderstanding about the GPL and OSS. Most of that's down to genuine ignorance that there are issues to understand rather than anything intentional. It's not surprising therefore that there too misunderstandings here. On the whole, and unless we get down to the nitty-gritty, it's pretty straight foward.

In particular, GPL software appears in many many places, both comemrical and non-commerical, embedded in hardware and as standalone software. It has been said very recently that much of Linux's success was because the GPL "forced" companies, especially the big name ones, to contribute back their efforts. By contrast, the demise of the NetBSD project was because they _didn't_ use this style of license. The RO licenses is different again, and I can't be sure how it might play out, nor what might constitute "commerical", since there are a number of ways to play that game.

mrchocky on 30/09/06 10:48PM [ Reply | Permalink | Report ]

Well I don't know what happened to my post but here goes again.

Well just as I was starting to feel positive along comes a dose of reality courtesy of ROL. No well this is positive news and we look forward to looking at the options to join forces and move RISC OS foward instead we get this an attributed comment along the lines of this changes nothing. Where is the vision to say we look forward to jointly developing code to expand the ability of Unicode aand to add to Select. Well I'm within a whisker of saying forget it. I've purchased RISCOS4 more than once, been a Select subscriber and Foundation subscriber and even though I've had an Iyonix for 3 + years I've still subscribed to support the market for nil return. I might as well have saved the £250 and got an A9 because at least then I would have had something. No doubt I can expect a mention in the next ROL newsletter but sometimes you just have to say what you believe in and at this moment my belief in ROL and Select is at an all time low.

bluenose on 30/09/06 11:01PM [ Reply | Permalink | Report ]

Browse and Fresco both have six letters in their name. That plus a bottle of Argentinian red wine causes one to write crap........

mripley on 30/09/06 11:09PM [ Reply | Permalink | Report ]

adh1003: "IMHO, it will be entirely possible to build a box using open RISC OS components and sell it. You would have to pay a unit royalty to Castle, but that would be your one and only notable restriction."

It's worth noting here that the one and only notable restriction, whilst it may sound like "just a few quid on the retail price" to some people, is less a financial restriction and more one of control. One major difference between some GPL+proprietary dual-licensing scheme and this one is that in the case of a proprietary licence being denied to a vendor, any released software must become perpetually non-commercial; thus downstream recipients of such code have to be careful not to profit from providing the code to others.

steelpillow: "This shared source license looks at first glance very similar to the kind of thing that MySQL, KDE and so on do, I'm not sure why some people think it is different."

It is different because even permissive open source licences do not distinguish between different fields of endeavour. And even the GPL focuses on the social contract aspect of keeping the source available to everyone, not on insisting on non-commercial usage, but that's because the FSF isn't about preventing people from making money: it's about preventing people from exploiting end-users by withholding the sources.

guestx on 30/09/06 11:09PM [ Reply | Permalink | Report ]

In reply to Jaffa: "RISC OS 6, free for end-users, running on RiscPCs is an appealing prospect, given the RPC still makes up the majority of current hardware."

I wouldn't be too sure that RPC makes up the majority. There could be many, many, many more people who, just like I, dumped the anomaly in favour of the Iyonix

;-)

cheers!

Ike on 01/10/06 02:02AM [ Reply | Permalink | Report ]

Guy Inchbald wrote: "This shared source license looks at first glance very similar to the kind of thing that MySQL, KDE and so on do, I'm not sure why some people think it is different."

KDE isn't licensed in the same way; you're thinking of Qt which is available under the GPL and a commercial license.

davidb on 01/10/06 03:26AM [ Reply | Permalink | Report ]

In reply to mrchocky: " That comment was made on csa.programmer, and you've replied here, with no indication that it was so, nor where to look for what else I said if others cared about the context. If you care to reply there, I will be happy to respond."

Well, I did reply to something I read in this drobe article in the "Reactions" part, the paragraph starting with "GCCSDK developer and Firefox porter Peter Naulls said". How am I to know that that was taken from some other place...

hzn on 01/10/06 07:47AM [ Reply | Permalink | Report ]

To Andrew Rawnsley in Reactions: "Unfortunately, most of the items listed have better versions already in existence as part of Adjust, so really all it does is accentuate the OS split, because the public only have source access to 'out-dated' versions of things."

For one ROL could offer quite a bit of their Select things for RISC OS 5 - not everything due to the kernel differences, as PM states. PM even showed the odd Select feature running on top of RISC OS at the odd show but I have the impression that PM has no intention of offering Select features and thus e.g. enhanced base apps like Paint etc. Furthermore until now it doesn't look like an IYONIX Select will happen by in near future. So the OS split is probably something we have to live with for some time thus why not enhace the IYONIX version?

hzn on 01/10/06 07:57AM [ Reply | Permalink | Report ]

Does anyone else think that Castle "opensourcing" essentially just the ROM apps (!Paint, !Draw etc.) is just their way to get the RO5 apps up to the spec of Select's apps, without paying for their development, and at the same time stopping ROL from benefitting either?

It's not as if their going to release the real meat of RISC OS (kernel, WIMP etc.) so that you could effectively compile your own RISC OS.

As for the Qt (not KDE) dual-licensing, it has for years been a bad bit of P.R. for Trolltech, until they finally gave in and released a GPL version of Qt4 for Windows last year, although it's still a bit cut-down (no XP styles, doesn't compile under VisualStudio etc.)

simo on 01/10/06 10:49AM [ Reply | Permalink | Report ]

Very few opensourcing of RISC OS so far apart from Unicode and the SharedCLibrary. It seems they want to only opensource peripheral components. With that in mind I think there are a few more components that would be interesting to propvide (if they have the rights to them): !Java and CDFS related components.

andretim on 01/10/06 11:08AM [ Reply | Permalink | Report ]

Thanks to all for the enlightenments and corrections (sorry, when I typed 'KDE' my brain was thinking 'Trolltech').

I think that we should see the present licensing move as step two (step one was open-sourcing the printer stuff) on a long and agonising learning curve to a new way of doing business. It is bound to have its faults, but then, that's what step 3 is for....

steelpillow on 01/10/06 12:19AM [ Reply | Permalink | Report ]

>It's not as if their going to release the real meat of RISC OS ( kernel, WIMP etc.) so that you could effectively compile your own RISC OS

Is that your conjecture? The FAQ on the site clearly states that the release will be phased and the software components listed are merely the first phase, to test the waters. They are the easiest for people to build and work on too.

adrianl on 01/10/06 1:03PM [ Reply | Permalink | Report ]

"As for the Qt (not KDE) dual-licensing, it has for years been a bad bit of P.R. for Trolltech, until they finally gave in and released a GPL version of Qt4 for Windows last year, although it's still a bit cut-down (no XP styles, doesn't compile under VisualStudio etc.)"

This is a rather narrow view of events, and since KDE very much relies upon Qt, you cannot separate the two in licensing terms. Try this for starters:

[link]

Qt was subsequently dual-licensed as GPL, as we now see from its appearance in Debian and other major distributions since 1998.

The Wikipedia entry talks about the Windows effort:

[link]

What I fear, and what seems likely, due to the "fields of endeavour" bit mentioned above, is that the GPL will be incompatible with the Castle Shared Source license. This is going to create plenty of headaches.

mrchocky on 01/10/06 1:50PM [ Reply | Permalink | Report ]

I think that what keeps a lot of people involved with RISC OS, is the "what happens next" factor. Open source should provide plenty more entertainment for them (us?). The nasty thing to say would be "too little, too late", had Acorn made this announcement in 1992 things would have been very different. Having dealt with them at that time, I can imagine their reaction to the idea. A few predictions: Rival versions of Paint, Draw etc. Rival "distros", ie. rival versions of RISC OS Entities that construct "distros" i.e. sets of RISC OS modules and applications that are consistent Entities that control and co-ordinate development of apps and modules More heartache for application developers ensuring compatibility with different modifications Huge row the first time ROSL are inspired by something someone adds to RISC OS 5 Huge row as people are inspired by Select/Adjust to add to RISC OS 5 RISC OS 5 for old machines e.g. RISC PC Paint, Draw etc. for Windows/Linux Nothing much happens because there are too few programmers left Martin Wuerthner adds GDraw functionality (anti aliasing etc.) to Draw module Increase in entropy in RISC OS space

DavidPilling on 01/10/06 2:25PM [ Reply | Permalink | Report ]

Simon John wrote: "As for the Qt (not KDE) dual-licensing, it has for years been a bad bit of P.R. for Trolltech, until they finally gave in and released a GPL version of Qt4 for Windows last year, although it's still a bit cut-down (no XP styles, doesn't compile under VisualStudio etc.)"

Yes, you have to have a commercial license to get the Visual Studio Integration features:

[link]

I don't use Windows XP, so I'm surprised to hear you say that about the Windows XP style. I'd be interested to hear about your experiences with the open source version on Windows, but you should probably contact me privately rather than follow up here.

davidb on 01/10/06 3:06PM [ Reply | Permalink | Report ]

But David, almost all those things have already happened ;-)

Rival versions of paint, draw - the Select versions are quite a bit better Rival versions of RISC OS - RO5 vs RO4 Paint/Draw for Windows - VRPC/RPCEmu, CorelXara (well, that's pushing things) Nothing much happens because of too few programmers - that's already the case. etc, etc.

mrchocky on 01/10/06 3:18PM [ Reply | Permalink | Report ]

Mrchocky:

You could be right about the GPL vs. Castle license. The GPL2 vs. GPL3 issue is shaping up to further muddy the waters - GPL3 is very much about preventing hardware lockdown (Tivoisation) of software in embedded systems, and so is pretty much designed to be incompatible with licenses like Castle's. If the Linux community stick with GPL2, as seems likely, the applications communities (especially for hardware drivers) may do so too and the GPL3-specific problems may not in practice matter to us. (caveat: as has been noticed, my thoughts are not always sensible).

steelpillow on 01/10/06 3:24PM [ Reply | Permalink | Report ]

David Pilling wrote: "The nasty thing to say would be ''too little, too late'', had Acorn made this announcement in 1992 things would have been very different. Having dealt with them at that time, I can imagine their reaction to the idea."

It might have been an interesting proposition, even later in the 1990s, but I can't help feeling that those old attitudes were prevalent in certain circles at Acorn right up to the end.

"A few predictions: Rival versions of Paint, Draw etc. [...] Paint, Draw etc. for Windows/Linux"

That's already happened to a certain extent, at least for Draw.

davidb on 01/10/06 3:25PM [ Reply | Permalink | Report ]

In reply to David Pilling and Peter Naulls: "Nothing much happens because there are too few programmers left"

Well at least there is some chance that something may happen to RISC OS even if it is just the odd small thing to start with ... sounds like a positive chance compared to now ;-)

(Sorry, I couldn't resist)

hzn on 01/10/06 3:30PM [ Reply | Permalink | Report ]

I think this is is very encouraging. It means at the very least that developers can fix those anoying little bugs or add a missing bit of functionality and easily and have them merged into the distributed releases.

For larger projects, there is a load of potential (for example filecore/fileswitch 64 bit support for hard disc MP3 players). The problem here is that large projects such as these need a lot of expertise and time which may not be available.

I don't think it will cause a proliferation of incompatible versions of the OS, not least because there are insufficient people with the time and skills to do this. For the forseable future, I think the majority of users will stick to official releases as a guarentee of quality. Similarly, I would expect most developers to develop for the official releases and not rely on custom enhancements.

I hope ROOL provide CVS or SVN repositories for public projects, aiding and encouraging colaboration and guidelines on new APIs and the like to avoid incompatibilities.

RISC OS Ltd. could sulk and refuse to accept the new regime, or they may be able to take advantage of it. They may be able to add these much talked about APIs to RISC OS 5 to allow more of their extensions to run and sell to Iyonix users. It is up to them to take advantage of the possibilities.

jamesp on 01/10/06 3:40PM [ Reply | Permalink | Report ]

steelpillow: "You could be right about the GPL vs. Castle license."

That's certainly something which could really torpedo any larger scale development. If you can't combine LGPL/GPL-licensed code with the "in due course" licence of RISC OS "Open", then everyone will be looking around for permissively licensed code or writing stuff from scratch.

steelpillow: "The GPL2 vs. GPL3 issue is shaping up to further muddy the waters - GPL3 is very much about preventing hardware lockdown (Tivoisation) of software in embedded systems, and so is pretty much designed to be incompatible with licenses like Castle's. If the Linux community stick with GPL2, as seems likely, the applications communities (especially for hardware drivers) may do so too and the GPL3-specific problems may not in practice matter to us."

The GPL in its current form already affects the announced developments, as mrchocky pointed out. What the GPL 3 seeks to achieve is the continued guarantee of the various rights described in its predecessor, doing so by explicitly forbidding any violation of the spirit of the licence either by technical means or by the use of other legal instruments to distort the application of copyright law.

Think of the GPL 3 as the GPL 2 without the loopholes. Unlike the vocal Linux kernel developer minority who don't seem to understand this, I predict that the GPL 3 will be immensely successful, not least because it will incorporate compatibility with various Apache licences and other open source licences which are currently technically incompatible with the FSF licences.

guestx on 01/10/06 4:29PM [ Reply | Permalink | Report ]

In reply to jamesp: I think it won't matter whether ROL sulk or not, if RO5 gets ported to the RiscPC and A9, then we have a single version of RISC OS for all reasonable RISC OS computers (i.e. no Archies or anything), that is of course what we all want, and there will be very little incentive to buy Select/Adjust. I'm not sure whether the ROOL thing will work out, but it would have to go really quite horribly wrong to work out worse than the mess ROL has made.

thegman on 01/10/06 4:54PM [ Reply | Permalink | Report ]

David Pilling wrote>"Huge row the first time ROSL are inspired by something someone adds to RISC OS 5"

If RISC OS Ltd use the *source* as a basis for inclusion of new features into Select then presumably they'll be obliged to *pay* royalties for that use. If they implement their own variant - well then they do and there isn't a whole lot anyone can do about it.

I would point out, however, that ROL could see the direction Castle were taking with RO5 and seemed not inclined to "copy" the methodology of any of it (32bit Select still has the WimpSlot restriction and no HAL to name two examples). I don't believe ROL will copy anything - if for no other reason that the currently available material (Paint, Draw etc.,) they've already upgraded. RO 5 is playing catch-up, so it's unlikely that copying any changes would benefit ROL.

Later releases might be more significant (e.g., Kernel, Drivers etc.,) - but all that is likely to do is remove the original excuse ROL used for not releasing Select on Iyonix (lack of information) - but in the meantime RO5 may well have gained similar UI features to Select - in which case the demand from Iyonix users for Select will have dried up.

David Pilling wrote>"Huge row as people are inspired by Select/Adjust to add to RISC OS 5"

Simply copying a method of "doing things" is probably Ok. A complete "rip-off" by reverse engineering big chunks of RO4 Select *would not be*. Anyway ROL had a perfectly legit way of preventing this - *releasing RO Select on Iyonix*. They didn't - and it would be a bit rich if they *now* complain that somebody else might add Select style UI features to RISC OS 5.

As far as ROL is concerned the boat has left without them - that was their chance they missed it - tough.....

AMS on 01/10/06 4:58PM [ Reply | Permalink | Report ]

GuestX: "...is less a financial restriction and more one of control"

In my opinion, no, this is not the intention. If the software is under the GPL then the viral-like nature of that licence infects software developed bespoke for the device in question. NRE is often a major proportion of the cost when developing a new embedded device. Many companies do not like the idea of putting effort into a product only to have to release all code to it - it's akin to writing off the NRE cost. Moreover, upon releasing sources, any potentially cheaper manufacturer in the world (notably China) will find it easier to copy the hardware, tweak the firmware in source form and erode the market started by the original innovator.

Some companies do release products around GPL operating systems like Linux, but they often either have a heavy string of software patents protecting other aspects of their design in the US (c.f. Tivo) or near-software patents good enough to scrape through in the EU, the product may be not in the slightest bit innovative anyway (e.g. a router), or it may be that the company is very large and feels able to defend itself on (say) the grounds of copyright violation rather than just accept a copy of its hard work appearing in the marketplace. Some even just ignore the code release provisions of the licence and cross their fingers hoping nobody notices! To avoid endless circular debate here, though, I would point out that there are of course countless permutations of this and no "right" answer when describing usage cases.

The reason why a royalty is payable for commercial RISC OS [Open] users is partly because those users can keep their code or modifications to core RISC OS private. They pay for what amounts to exclusivity. It's about control in a way, as you said - but not Castle's control, as you indicated; I believe it's actually there to grant more control to the user of the code.

adh1003 on 01/10/06 6:09PM [ Reply | Permalink | Report ]

It is easy to understand some of the more cynical views expressed herein, but as for the intial release of components: We have to start somewhere. RISC OS is complex software with a complex history. When ROOL was first "noticed" (IIRC by its Companies House registration) there was a lot of speculation about what we'd be doing. Some people discussed legal issues with releasing RISC OS source code to the public and indentified, in part correctly and in part incorrectly, some of the legal hurdles that are present. Put in its most basic form, we can't just dump the whole CVS tree on a Web server and say "go for it".

The intention after the initial source release is to continue with more components. The ultimate goal would be to have a complete OS there - perhaps, for example, you could build yourself an Iyonix ROM. But that's putting the cart before the horse. We cannot promise being able to reach such a position at this stage, so we're not doing so. The list of components is what it is; hopefully there will be more, but even if this is all that ever gets released, surely it's better than nothing?

adh1003 on 01/10/06 6:15PM [ Reply | Permalink | Report ]

David Pilling:

Personally, I think "too little, too late" has a ring of truth, depending on your perspective. But "better late than never" has a lot of truth too and you have to bear in mind that the political and legal climate must be considered. IIRC Castle have previously stated clearly that they were not willing to release source code, but times change and it would be madness to ignore this and cling on to some tired mantra in ignorance of the world around you. Of course, some people have complained that Castle have changed their mind on the open (strictly, shared) source issue; just as some were complaining when they didn't change their minds. That's life.

As for RISC OS turning into multiple little distros, well, that might happen. We're unlikely to ever have the number of developers that something like Linux has, but even with a smaller number, divergent releases are a possibility. As a community, we have to consider:

* Are multiple distros in fact a healthy sign of a varied developer community? Bear in mind the RISC OS philosophy of forwards and backwards compatible APIs and its modular nature. A "distro" containing one set of modules by default might just mean a set of "*RMensure" lines in an application and softloading what you need on a distro containing another set of modules by default. Not much different from loading shared libraries on any other OS really. Application developers seem to cope.

* Can the community, if changing APIs, can maintain the high quality (on average, with obvious exceptions!) of the Acorn-designed APIs in the existing OS? Well designed interfaces with graceful fallback (or fall-forward) can help mitigate application development problems. Many RISC OS developers share the design approach/philosophy that underpins parts of RISC OS so they do produce modules with well conceived interfaces. If the developer community expands, it will be up to existing developers to provide new developers with support and documentation to help them achieve what we believe is a robust and considered design, while recognising that everybody learns from everybody else all the time. If we instead choose to turn inwards and ignore one another then the results are likely to be grim.

* Does the community wish to submit changes and improvements back to ROOL for inclusion in the master tree to try and avoid fragmentation? If people believe that divergent code is the way forward then that's what they'll do. If people believe that convergence is essential, then we've done everything we can to provide the support mechanism through a neutral third party. The community makes its choice.

adh1003 on 01/10/06 6:33PM [ Reply | Permalink | Report ]

guestx: "Think of the GPL 3 as the GPL 2 without the loopholes. Unlike the vocal Linux kernel developer minority who don't seem to understand this..."

This is a very partisan issue, with much heat on both sides. One ma's loophole is another man's lifeblood. For more on the debate. See for example the LWN discussion at: [link]

steelpillow on 01/10/06 7:16PM [ Reply | Permalink | Report ]

Apologies. That should be: [link]

steelpillow on 01/10/06 7:20PM [ Reply | Permalink | Report ]

Andrew Hodgkinson wrote: "To avoid endless circular debate here, though, I would point out that there are of course countless permutations of this and no ''right'' answer when describing usage cases."

Quite, but it would help to avoid circular debate if you could avoid using contentious terms like "viral" and "infect". :-(

davidb on 01/10/06 8:28PM [ Reply | Permalink | Report ]

Andrew Hodgkinson wrote: "The reason why a royalty is payable for commercial RISC OS [Open] users is partly because those users can keep their code or modifications to core RISC OS private. They pay for what amounts to exclusivity. It's about control in a way, as you said - but not Castle's control, as you indicated; I believe it's actually there to grant more control to the user of the code."

I agree that endless circular debate is unnecessary, however although the above is a laudable sentiment - and great from a marketing point of view - without the choice of dual licencing (which would require copyright assignment from community contributions, which may be unpalatable to some) commercial developers may have more *control*, but they have less *choice*. They can't not "pay for exclusivity" if they don't need it.

Still, I think ROOL's efforts on the whole are going to be better than nothing; it'll remain to be seen if it's revolutionary enough to revitalise the market, but at least it'll give people itching to fix bugs in Paint a chance to do so.

(If I was being cynical, I'd have subsituted "Paint" for "Edit" in there... ;-)

)

Jaffa on 01/10/06 8:44PM [ Reply | Permalink | Report ]

Ohh l,ots of people cry out for open source RISC OS, and just as it is made available, we get the GPL is this and GPL is that brigade coming out of the woodwork qande arguing about GPL rather than being positive about the OS at hand.

Look, it is simple........Castle are releasing the code for many to play with and they can feed back the details, however, the OS is castle's and theirs alone to decide.

Just be thankful, that it has arived and you now can hgave a look.

If you dont like their terms and conditions then it is simple. Don't touch it with a bargepole and get on with your life. Dont get on your high-horse and rant.

Or if you don't like the terms but really really feel for the OS, then bite a bit of pride and lend a hand. Otherwise do something else.

If you like the terms and conditions and can help then do so with a smile on your face.

I'm away to ask about porting my fav IDE to RISC OS now.

Cheers Bob

nijinsky on 02/10/06 09:48AM [ Reply | Permalink | Report ]

Bob Hartley wrote: "Ohh l,ots of people cry out for open source RISC OS, and just as it is made available, we get the GPL is this and GPL is that brigade coming out of the woodwork qande arguing about GPL rather than being positive about the OS at hand."

People are only clearing up the usual misconceptions about it that always seem to come up whenever the subject of licensing is raised. As usual, most of the negative rhetoric comes from the people who don't like the GPL for whatever reason. Something else you said applies to those people, too:

"If you dont like their terms and conditions then it is simple. Don't touch it with a bargepole and get on with your life. Dont get on your high-horse and rant."

davidb on 02/10/06 11:48AM [ Reply | Permalink | Report ]

Picking up Andrew Hodgkinson's points. I do think this is a positive step, and I didn't expect it to happen. Congratulations to those who have achieved this. Fragmentation is a possibility, and there could be incompatible changes made, who will decide between them? However I suspect there won't be much of a problem in reality. As a grumpy developer I hate different versions of RISC OS working in different ways - it is no fun at all having to do work for the sake of 10 people who have version X of RISC OS. RISC OS might have had a new generation of scanner drivers if I hadn't used up so much time with the Simtec/Castle USB API episode. Yes on reflection we have had a some things I predicted before with Select/Adjust. Two things might be different. Firstly Select was effectively a complete version of RISC OS. You didn't find someone using random components of Select on RISC OS 3.7 or the Iyonix. Secondly Select was a paid for upgrade which limited the numbers using it. There is an assumption there, open source == free software?

DavidPilling on 02/10/06 12:00AM [ Reply | Permalink | Report ]

In reply to andretim: "With that in mind I think there are a few more components that would be interesting to propvide (if they have the rights to them): !Java and CDFS related components."

!Java couldn't be open sourced, though things like the threading module could (it's a bit of an embarrassment, though - it was a proof-of-concept hack I knocked together in a weekend that was always supposed to be rewritten, but never did).

However, from a legal point of view, with suitable (free) licenses available from Sun, and if Castle/ROO were willing, it'd technically be possible to share the source with specific named individuals, if they were willing to develop it - but the end product would have to be free, otherwise you'd end up having to pay a large licence fee to Sun.

I considered this in '99, but it was too much effort. You'd not believe how much work passing the validation suite is. I took out a licence in my name (with E14's permission), which I suspect is still valid.

Piers on 02/10/06 1:00PM [ Reply | Permalink | Report ]

"Ohh l,ots of people cry out for open source RISC OS, and just as it is made available, we get the GPL is this and GPL is that brigade coming out of the woodwork qande arguing about GPL rather than being positive about the OS at hand. "

Couldn't be further from the truth. GPL is a pervasive license, and it's crucial to be sure how such sources from Castle might interact with it and other similar licenses, and how this is going to be perceived outside the RISC OS world. Do you care to be in a position of distributing software that breaks licenses - I think not.

mrchocky on 02/10/06 2:32PM [ Reply | Permalink | Report ]

"RISC OS Ltd. could sulk and refuse to accept the new regime, or they may be able to take advantage of it...It is up to them to take advantage of the possibilities."

If I were RISC OS Ltd I'd be interested in becoming the defacto distribution. Perhaps they could go down a Mandriva-style route where official paid-up (40 quid a year? Count me in) members of the "ROL" club get 12 month advance access to their distribution as well as some extra support / newsletters / niceties. (The first release could largely be comprised of Select). After that period a new version is (hopefully) released and the older version submitted as open source / free download. Possible? I'm sure they'd hate the idea but i'm trying to think of a role for ROL and all their experience and Select development in this potential new exciting RISC OS world.

Ginger2 on 02/10/06 3:51PM [ Reply | Permalink | Report ]

I am very happy about this annoucement. I must admit I still need to get to grips with the licensing aspect of this though.

The thought of getting some real development of RISC OS that 'can' be poured back into a common source is a very good step forward.

I wonder how long it will be, before we can then get artists developing new system wide icons & audio sounds for errors etc.

This move just opens up a whole new dimension to revitalising RISC OS.

Well Done to all those involved - and thank you.

nx

nx on 02/10/06 4:45PM [ Reply | Permalink | Report ]

In Reply to Stephen Gadd

Your suggestion of a ROL club and how ROL could handle things is exactly the sort of thing I would expect ROL to come up with if they really thought things through. The fact that I've paid £250 for the existing ROL club for the past 2+ years with no returns, yet, would influence me to some degree and I'd certainly want some better assurances for a Iyonix element to the ROL Club. I do hope that ROL look at the options and come up with something that fits in with their business model but enables all those who have modern machines to be included. Select in the early days was great and I'm one of those who miss some of the features from it BUT RISCOS5 has some of it's own and a combined paid for and supported release would be the best for all the market. So I'd say to ROL to grasp the vision and grab a further slice of the market.

bluenose on 02/10/06 6:49PM [ Reply | Permalink | Report ]

adh1003: "The reason why a royalty is payable for commercial RISC OS [Open] users is partly because those users can keep their code or modifications to core RISC OS private. They pay for what amounts to exclusivity. It's about control in a way, as you said - but not Castle's control, as you indicated; I believe it's actually there to grant more control to the user of the code."

So what would be the problem with a choice of the GPL or some closed source licence? That someone could still sell GPL-licensed code and Castle wouldn't see any of that money?

guestx on 02/10/06 7:15PM [ Reply | Permalink | Report ]

David Pilling This item has attracted a lot of attention from David Pilling. Now I was under the impression that David had switched his attentions to Windows development. The fact that he is following this item raises some questions in my mind. 1) Would the move to Open Source cause David to come back to the Risc OS field. 2) Would he consider doing open source developments i.e. for free?

Jwoody on 02/10/06 8:08PM [ Reply | Permalink | Report ]

James Woodland... I'm here to see what happened to an old friend. If open source re-ignited interest in the desktop then yes I might come back, I don't think it will though. I believe the aim of open source is to boost embedded applications - so if I'm looking for a platform to implement my toast making algorithm on, I might now pick RISC OS. Would I contribute? Had the open source appeared five years ago then I would have done. Today I'm not in touch with what needs doing. I have quite a lot on my plate as it is. Never say never, but I can't see an immediate return to RISC OS. I would be open to contributing existing source code - if there was anything useful, thinking components or bits of code here not applications.

DavidPilling on 02/10/06 10:12PM [ Reply | Permalink | Report ]

A couple of additional entries have been added to the shared source FAQ on our site which cover the following questions:

Why didn't we select the GPL? Why did you select those components to open first? What tools do I need to build the sources?

Hope that helps.

riscosopen on 02/10/06 11:27PM [ Reply | Permalink | Report ]

In reply to quite a few here: What is amazing still is that instead of looking forward to the open sourcing and seeing it as a chance for the better an amazing part of this discussion looks into the licensing issue. I am sure that the ones behind the open sourceing did spend some time on thinking about this issue and thus came to the decision taken for good reasons. I kind-of have the feeling that some posters here might be tempted upon getting a present for birthday or X-mas they first of all start to ask if they may sell it on eBay if they don't like it. What a strange world...

In reply to Ginger2: "If I were RISCOS Ltd I'd be interested in becoming the defacto distribution. Perhaps they could go down a Mandriva-style route where official paid-up (40 quid a year? Count me in) members of the "ROL" club get 12 month advance access to their distribution as well as some extra support / newsletters / niceties." You sure that 12 months is enough considering that the last upgrade for Select took more than twice that time and is still not out ;-)

In reply to riscosopen: Thanks for the additional FAQ entries as well as your other work so far!

hzn on 03/10/06 09:52AM [ Reply | Permalink | Report ]

"You sure that 12 months is enough considering that the last upgrade for Select took more than twice that time and is still not out "

My dreams continue...

I'd recommend halting Select 4 development and putting their efforts into an intitial pay-for distro release (comprised of mostly Select 3 components) for Christmas/New Year. Select subscribers automatically get a copy of this inititial release (this could be a username and password for the members area of the website), non-Select people are invited to join the club (this includes me, where they get a welcome pack and install CD). I'd advocate a nice new spangly name for this distribution (and a new website!). After this initial release they can then concentrate on <newname> 2007 - which can include lots of the Select 4 components plus new work. This gives them time to finish what would have been Select 4. When this happens the initial release becomes available as a free download / source code.

Of course, now that the source code is available, there's potential for anyone to form such a company (just so long as ROOL licences aren't prohibitive!) and offer added-value features and services to the base open package. I just think ROL are in a great position to take the lead and gain some new respect.

I wonder if they're reading this thread....

Ginger2 on 03/10/06 12:08AM [ Reply | Permalink | Report ]

Stephen Gadd>"If I were RISCOS Ltd I'd be interested in becoming the defacto distribution. "

Why? What exactly *would* ROL bring to the party - they have shown a *complete* disinterest in Iyonix users. Additionally if ROL were to *charge* for any of the source *or* try to close it and include it in Select then they'd have to pay *a license fee* for such use.

I'd be happier leaving it in the capable hands of Castle or ROOL rather than RISC OS Ltd - besides as others have pointed out ROL are *already* obliged to their existing subscribers who have paid for goods yet haven't received them yet.

AMS on 03/10/06 1:40PM [ Reply | Permalink | Report ]

If you are interested in helping out with the RISC OS shared source project, take a look at the following page:

[link]

riscosopen on 03/10/06 3:11PM [ Reply | Permalink | Report ]

AMS:

But Stephen didn't say that *he* wanted ROL to become the defacto distribution; he said it would make sense for ROL to want that. He's probably right, but since most of ROL's other decisions don't make any sense, I don't think that they're likely to get involved here.

And I don't think ROL (which is to say PM, who runs it) is interested in delivering anything to anyone. They/he only seem(s) to be interested in prolonging the OS fork and making life difficult.

chrisj on 03/10/06 3:27PM [ Reply | Permalink | Report ]

hzn: "What is amazing still is that instead of looking forward to the open sourcing and seeing it as a chance for the better an amazing part of this discussion looks into the licensing issue"

Since there are few other details, and no actual source release yet, there is little else to discuss. Besides, as I said, the license remains critical - it very much affects what long time RISC OS open source advocates like myself can do with it (although I have slightly wider considerations than most RISC OS enthusiasts) and who already need to have a good understanding to ensure their activities remain above board. The GCCSDK project has gone to considerable lengths in this regard, and will need to continue to do so.

mrchocky on 03/10/06 3:27PM [ Reply | Permalink | Report ]

@hzn: The reason why people discuss the GPL compatability of the ROOL licence, is because it would be extremely helpful if developers could make legal use of the vast amounts of GPLed code available on the net to enhance RISC OS components.

JGZimmerle on 03/10/06 6:11PM [ Reply | Permalink | Report ]

What are the chances of getting the BBC Micro OS released as open source too?

(not a priority I guess, but...)

murkle on 03/10/06 7:02PM [ Reply | Permalink | Report ]

The onel thing that no one's realised yet is the possible change this makes for emulators. If it's possible to legally distribute a version of RISC OS - say, a cut down RISC OS 5 version (or perhaps the Pace RISC OS 4 version) for RiscPC-like emulated hardware, that could be a big boost to getting people to try RISC OS. Using it wtih GPL emulators like RPCEmu would be no problem, since it's just data it uses. (although RPCEmu would be classified as "contrib" by Linux distributions like Debian)

mrchocky on 03/10/06 9:52PM [ Reply | Permalink | Report ]

riscosopen: "Why didn't we select the GPL?"

Thanks for updating your site. However, the answer doesn't really address the dual-licensing issue (the choice of GPL or closed source) since it compares the GPL to closed source licences, whereas it should be standing up for the shared source "in due course" licence instead. We already know that paranoid device developers don't want to release their sources even, in some cases, when they've used GPL-licensed code. The answer doesn't address licence incompatibility or show that the "in due course" licence gives any particular benefits that outweigh the major downside of not being able to combine the code with GPL-licensed works.

As for the notion of relocatable modules, couldn't Castle just state, as the copyright holder, that they do not consider relocatable modules to be derived works of the operating system? I'm sure Castle managed to persuade various other copyright holders using a similar interpretation of the GPL when they squeaked through the alleged HAL "GPL violation" case. This time they actually have the authority to make that particular call, however.

guestx on 03/10/06 11:21PM [ Reply | Permalink | Report ]

I too would like some answers to these questions. With respect, it doesn't seem to me that this has been fully thought through. My preference would remain for a dual licensing system (such as many named in examples above) with a GPL-compatible license alongside the commercial one. And I would again stress that much of the popularity of Linux has come about precisely because of the GPL.

mrchocky on 04/10/06 00:45AM [ Reply | Permalink | Report ]

Ginger2: IMO ROL should finish Select 4 first (after all, people paid for it 2 years ago), open source suitable elements of the OS which need work (particularly low level components) and concentrate on selling a premium distro package like you suggested. PM has stated one reason why he doesn't like Open/Shared Source is the possibility of multiple versions of components and multiple distros coming about. This could be controlled by putting in place suitable measures to co-ordinate community development and encourage convergence, and/or by the type of license selected - although the latter, if more restrictive, could backfire and cause some developers to ignore RO4 altogether. Even so I can't see how that would be much worse than the current snail's pace closed-source development. Paul talked about taking risks with Adjust32, I think this is a more crucial one.

AMS: "Why? What exactly would ROL bring to the party - they have shown a complete disinterest in Iyonix users." I don't think ROL can afford to ignore Iyonix users any longer. Limiting themselves to only 26-bit hardware (which, apart from the R7500, is entirely out of production) and the A9 is making a tiny market even smaller. What would make things really interesting is if Castle released a new mobo before Iyonix Select was finished, then Paul's company would be even more overstretched.

chrisj: "They/he only seem(s) to be interested in prolonging the OS fork and making life difficult." Unfortunately with a dwindling customer base and limited scope of resources, one wonders if that's all ROL are able to achieve.

timephoenix on 04/10/06 05:27AM [ Reply | Permalink | Report ]

Peter:

"With respect, it doesn't seem to me that this has been fully thought through."

With respect, I think it's probably been thought through to a greater depth that you believe, most likely by all the relevant people involved with ROOL and Castle, along with legal bods who understand this sort of thing. Suitable discussions will have been taking place, in all likelihood, since before ROOL was even formed, let alone brought to everyone's attention on Drobe - probably less formal and with less well defined plans in those early stages, but you can bet your life that a great deal of thought and careful study of the possible licencing arrangements (and indeed licences such as GPL) has taken place in the intervening period.

VinceH on 04/10/06 09:10AM [ Reply | Permalink | Report ]

Er yes, but again, the inconsistencies in the answers remain. IANAL, but deal with this stuff all the time, and understand precisely the lengths you could or could not go to. It's all very well saying "well, they _must_ know better than you because they have legal bods", etc, etc. Let's stick to the issues at hand rather than trying to one up.

mrchocky on 04/10/06 09:26AM [ Reply | Permalink | Report ]

I'm not "trying to one up" but thanks, al the same, for ascribing an intent to me that wasn't there.

VinceH on 04/10/06 09:45AM [ Reply | Permalink | Report ]

Great, so we're now at the level of reading personal insults into posts. Let's try and stick to the issues at hand.

mrchocky on 04/10/06 3:21PM [ Reply | Permalink | Report ]

Peter: of course someone has thought about the emulation situation, as you can find out when reading all those well-thought-out comments by developers.

Oh btw, thank you Chris for leaving out a significant part of my statement without even indicating it. Is that the standard of journalism we can expect from Drobe nowadays?

hubersn on 04/10/06 3:42PM [ Reply | Permalink | Report ]

VinceH: "With respect, I think it's probably been thought through to a greater depth that you believe, most likely by all the relevant people involved with ROOL and Castle, along with legal bods who understand this sort of thing."

Well, this is the peanut gallery, and if the "relevant people" can't even muster a decent defence of their mythical shared source licence to us, then I think you put too much faith in those "legal bods" who, it would seem, clearly have to work a bit harder to "understand this sort of thing" if they think that yet another incompatible pseudo-open licence will somehow catalyze massive public interest in their investment.

guestx on 04/10/06 8:24PM [ Reply | Permalink | Report ]

GuestX:

It's more a case of NOT having faith in what is nothing more than a speculative dismissal of an unknown than of having faith in that unknown. You are condemning them based on what you think might be the case. I am reserving judgement.

Peter:

I don't consider the "trying to one up" as an insult, per se, but given the sentence that contained it, it is clearly what you are accusing me of doing in my reply to you. If you don't want me to read it into your post, try not putting it there in the first place.

VinceH on 04/10/06 8:38PM [ Reply | Permalink | Report ]

Slight correction to my post - I am reserving judgement, but for the moment I believe they should be given the benefit of the doubt, rather than dismissed based on unknown factors, or something like that. I can't remember what I was actually going to write - it was more than two minutes ago.

VinceH on 04/10/06 8:43PM [ Reply | Permalink | Report ]

steffen:

I imagine this will spark more conspiracy theories rather than solving anything, but for the record, below is Steffen's unedited email as it arrived. I hope Steffen will agree that he is not in any way misrepresented. The question about ROL was removed because it looked silly coming straight after the ROL quote from Paul, but of course when I emailed Steffen he had not seen Paul's response.

-------

Date: Sat, 30 Sep 2006 02:23:55 +0200 From: Steffen Huber To: diodesign@diodesign.co.uk Subject: Re: RISC OS shared source announcement

I really welcome this move. To me, it is the first positive piece of news for our market since the launch of the IYONIX pc. It is probably the only way to keep RISC OS alive, and it has the potential to interest many ex RISC OS developers.

It is of course too early to consider all consequences - how will RISCOS Ltd. be affected, what exactly is that "per unit royalty fee", would this "shared licence" allow royalty free shipping of a freeware emulator complete with RISC OS, does e.g. selling support for an otherwise free product make it a commercial product...the coming discussions will keep us surely busy for a while ;-)

Steffen Huber, hubersn Software

------

"Thank you Chris for leaving out a significant part of my statement without even indicating it. Is that the standard of journalism we can expect from Drobe nowadays?" is now the 67th quote for the top-right-hand corner.

diomus on 04/10/06 10:08PM [ Reply | Permalink | Report ]

VinceH: "You are condemning them based on what you think might be the case."

No, I'm questioning them on the basis of what they've said. As mrchocky pointed out, rolling your own licence which will be GPL-incompatible, based on what has already been stated about it, raises important questions about their expectations and motivations.

We obviously await the full text, but given that a "free of charge" non-commercial stipulation is obviously central to the licence - one wouldn't prominently and repeatedly mention things like this if they're just minor details yet to be thrashed out - we know straight away that there will be serious incompatibilities with open source and Free Software licences - it's even admitted in the FAQ that "the shared source RISC OS project is not open source for the simple reason that it makes a distinction between people who want to use or access the RISC OS sources for non-commercial purposes and those who want to use them for commercial purposes".

So, what are we left to expect? Possibly the only remaining thing of interest is the reasoning behind it all.

guestx on 04/10/06 10:46PM [ Reply | Permalink | Report ]

Please login before posting a comment. Use the form on the right to do so or create a free account.

Create a new account
Forgot your password?

Search this website

This week's poll

Recent developments have left me feeling...

Assured ROS will appear on new hardware		55%
Pleased OS desktop features are being developed		10%
ROL and ROOL will eventually kiss and make up		5%
App developments are critical		10%
Dave Holden sleeps easy at night		19%

Discuss this | Archives