An idea to make RISC OS Open Source

Rick:

Whoa. You must have some really old tech if you actually need to open it up to change the firmware.

Not really. I bought an eReader a while ago that boots from a micro-SD card. Quite hackable and fairly cheap, too. I installed a recent version of Debian on it, built a stripped down set of Qt libraries to run on it, and wrote some applications in Python to get it to do what I wanted it to do.

Well, I think I pretty much got what I expected from this thread. I’m not about to invest much more time in trying to defend the principle of openness on consumer devices because it’s off-topic here. I don’t have an preferred choice of license for RISC OS, though I think chasing up rights holders as soon as possible is a worthwhile thing to do in any case. I wrote a piece of software for the OS in 1996 for which the complete work has three copyright holders, and chasing up the others is already a difficult task.

Personally I think you mix up the benefits of RISC OS being truly open-source with how you justify your day-job as software engineers.

Oh, he must think everybody gets paid to develop the OS. How cute.

GPL gives you rights which cannot be taken back; that’s the whole point.

It is fascinating to see how people support the GPL and it’s stance on DRM. They don’t realise that the GPL and its threat of legal action if you don’t comply within the fuzzy vague constraints of the licence itself is, in essence, much the same thing. It wouldn’t kill you to read through the CDDL text.

Earlier, I stated that GPL was only compatible with itself. I was wrong. It seems as if GPL v3 is not directly compatible with GPL v2, to the extent that you may not in fact satisfy GPL v3 if you mix in GPL v2 (only) code. http://gplv3.fsf.org/wiki/index.php/Compatible_licenses#GPLv3-incompatible_licenses

the freedom to use the software for any purpose, the freedom to change the software to suit your needs, the freedom to share the software with your friends and neighbors, and the freedom to share the changes you make.

Answers: except commercial we can, we can, we can, and we can.

There is no commercial value in RISC OS anyway

And there you have killed your own argument. It doesn’t matter that we can’t sell RISC OS or that Castle can, because nobody is buying. So we can thus ignore the OEM rubbish in the licence. The rest? Reads like an open source licence.

it’s a fringe OS which is only of interest to hobbyists; the castle license restricts only those who have no commercial interest in an OS which has, let’s face it, zero commercial value.

Translation: wah!

All of a sudden abandon-ware seems quite an attractive ideal

Abandonware? It isn’t a legal concept, it is a bunch of people deliberately disregarding licences and copyright because they think the software is too old for anybody to bother prosecuting over. You feel that abandonware is an attractive ideal and you want RISC OS to fit into your concept and definition of free (and xxxx what others might think if they don’t agree). You’re quite the freetard, aren’t you, eh?

Goodbye then. Enjoy your Amiga. Back in the day myself and a friend loved playing Cannon Fodder.

And there the Castle “shared” licence fails.

By the way, just to nit-pick, we (lowly coders) may think of RISC OS as having a practically open source licence, and for non-commercial use, it is as good as…but nowhere actually tries to claim that it is open source.

RISC OS Open may be so-called because the source code is now “open” and accessible (using common English rather than the specific OSI definition), something we didn’t think was possible until it started happening.

The licence information page says this:
The Shared Source RISC OS project, which can be found in the “castle” section of our CVS repository, publishes software and sources under Castle Technology Ltd’s dual-license scheme. More information about Castle’s license terms can be found here. A copy of the Castle non-commercial license can be downloaded here (PDF format).

I can only assume that before launching into discussions regarding licences and such, nobody actually bothered to read ROOLs own position on the matter? https://www.riscosopen.org/content/documents/ssfaq

“Oh, he must think everybody gets paid to develop the OS. How cute.” – way to make a statement out of the blue with no grounding. My point was that only musicians would argue that limiting the public’s access to art is a good thing in the same manner than only a professional software engineer would argue that limiting the public’s access to code is a good thing.

As for abandonware; it might not be a legal concept, but, one could argue that the continued involvement of active commercial entities in RISC OS has restricted hobbyists in many ways: "Come try RISC OS on the Raspberry Pi it’s free! Oh, you’d like to change the SD Card image? Well, that’s 40 quid – sorry about that, but “excuse here” is the reason we don’t use GCC". Want to run an emulator for a A3000 and play Elite? Well, we can’t help you with ROMs, that’s stealing.

“And there you have killed your own argument. It doesn’t matter that we can’t sell RISC OS or that Castle can, because nobody is buying. So we can thus ignore the OEM rubbish in the licence. The rest? Reads like an open source licence.” – nope; the license is restrictive – if there is no commercial value why agree with a license that restricts you only because of some historic value assumed by a, basically defunct, company?

Having said that, Rick, I have read ROOLs standpoint on licensing; however, if we return to the original idea which caused me to open this topic in the first place, I think we could do better and be really open-source. Seems that only David B’s post above seems to recognise this. Better to do something about it now than wait until it’s even harder to achieve. I think we could crowd-source a free RISC OS; but it requires the nay-sayers to admit that there is a problem with the Castle licence in the first place.

x

I’m persisting because I think it’s a good idea for RISC OS to be free. However if the main-stay of the OS developers (whom I’m presuming are well represented on this forum) disagree, then it’s a bit of a non-starter. If you go back and read the OP you will see exactly what I was asking:

“So, here is an idea. Why doesn’t someone at ROOL who knows the person/people behind Castle ask them how much money they want for RISC OS under the proviso that, if the licence were transferred to ROOL (and I mean code, trademarks, etc.) it would be released under the GPL licence. We could then start a kickstarter (or use any of a number of fund raising websites) to raise the money and buy RISC OS.”

I admit I am presuming that someone lurking around here knows the 3 people who are associated with Castle? The names are on Wikipedia, but I didn’t have much luck tracing them via Linkedin.

Yes, this has spiralled down into a 2 pronged conversation (i.e. “GPL is bad!” and “Castle’s shared source license is acceptable even if it’s not truly open”) and the original intent seems to have been ignored.

The overall level of negativity and naysaying on this topic has been surprisingly high. If you feel that there are no naysayers here then perhaps the dictionary definition will help:

“someone who says something is not possible, is not good, or will fail”

There has been way too much of that and not enough “how could we make it happen?”. When you have people posting “wah!” in 200pt purple text you can understand why the status quo is what it is.

So, in a way you’re right; before I was motivated and now I am not.

x

way to make a statement out of the blue with no grounding.

I really like how you continually cherry pick your quotes and still persist in telling us the GPL is the best licence despite several of us giving you actual technical reasons why it would be a bad idea for RISC OS. Even something like the CDDL might work, but the GPL absolutely won’t. This isn’t anti-idealism or me being obstinate, it is reality.

As for abandonware; it might not be a legal concept, but, one could argue that the continued involvement of active commercial entities in RISC OS has restricted hobbyists in many ways:

As opposed to the alternative being that it was an operating system that some people used to use back in the ninties bit! you know! you might find some ROM images online but that’s about it…

You mentioned the Amiga before. Well, AmigaOS is still around. It is closed source, and getting a copy will set you back over a hundred euros. http://amigakit.leamancomputing.com/catalog/advanced_search_result.php?keywords=Amigos

Want to run an emulator for a A3000 and play Elite? Well, we can’t help you with ROMs, that’s stealing.

Yes. It is. Funny how you want us to all swoon at the sight of your preferred software licence (which depends upon copyright in order to be effective), yet you don’t seem to have so much concern over the copyright of older things – hence putting “stealing” in bold, as if it is a bad thing that you can’t just take the ROM images and use them as you see fit.

think we could do better and be really open-source.

To what point? If OEM is a pointless concept, then an OEM clause will not be of consequence. So what, other than idealism, will changing the licence offer over that which we already have?

There has been way too much of that and not enough “how could we make it happen?”.

And from you, there has been a stubborn push to match what you think is the best idea while disregarding the reasons why it will not work.
This time I’ll keep my summary short:

• Castle does not own all of the code; changing the licence of those bits cannot affect the rest
• The GPL is a slimy insidious landgrab with many restrictions and numerous ill-defined concepts underneath all of the hyperbole. To apply it to RISC OS would be a disaster.
• It would be a disaster because alternative licences cannot co-exist with GPL in statically linked code
• Asides from commercial use, how does changing the licence offer us any actual benefit over that which we have now?

When you have people posting “wah!” in 200pt purple text you can understand why the status quo is what it is.

128pt and magenta.
You insult the OS, you insult the company that paid for it and gave it to us for free, you insult the people developing, and you seem to imply that our presence here affects your ability to just use old stuff as you wish without concern…… What else do you expect me to say?

Here is the bit where they deny me the right to make money:
“1.2.3 You shall ensure that each Derivative Work distributed by You is distributed and licensed in its entirety at no charge”

Sorry, but that doesn’t answer my actual question. Please give me an example. In particular deal with an example akin to: You write a module which you say can be freely added to the OS. But which you reserve the right to use yourself to “make money” in some way. e.g. by then offerring your own ‘extended’ version to your own paying customers as a 3rd party offerring.

Look at it another way. If I write some code that goes into the Linux kernel/modules, how would I then make money from it? Only by offerring an ‘extra’ either as training/documentation or by offerring and ‘improved’ commercial version for those who would be happy to pay.

I can see that Castle might make money from RO contributions. But how are they preventing you from also doing so if you make an ‘improved’ version that you sell yourself? I’m far from clear what you’re on about in practice rather than resorting to lawyerese that doesn’t clarify the point I’m asking about.

TBH Personally – linux or RO – I’m far more interested in making useful contributions that aid the community of users. Making money isn’t my real concern here. But I can quite understand why some will wish to. Be they Castle, or some of the contributors to linux like Red Hat. The details differ, but if your concern is ‘making money’ it remains possible. Indeed paid for user software is still alive for RO.

Jim

Rick, there are so many contradictory statements in your post that I don’t even really know where to start with it.

So here’s a short summary:

You hate GPL licences; I get it – and I disagree wholeheartedly.

Accusing me of doing exactly what you are, in fact, doing yourself in your posts is a little rich.

If you feel that you have been insulted, then I apologise. Although I would state that holding an opposing view is not the same as being insulting.

The Shared Source licence is not an open-source licence and it is unnecessarily restrictive (whatever your view on GPL). Changing to a GPL licence would place RISC OS in a better position in my opinion. Yes, it’s technically challenging but then so is software in general. It’s certainly not insurmountable.

Not contributed since the start of the thread as I was away across the country for a day (In a pub with friends with common volunteering ideals)

So (as suggested), I will have a stab at some application development instead which I can GPL. Shame though, since I have a lot of experience in low-level USB development which could have been a more beneficial contribution.

Seems reasonable use of your time. Perhaps work on RPCEmu would be attractive to you.
Various aspects are less than perfect and a good reliable free emulator to run the current OS release might attract a few more users.

Why doesn’t someone at ROOL who knows the person/people behind Castle ask them how much money they want for RISC OS under the proviso that, if the licence were transferred to ROOL (and I mean code, trademarks, etc.) it would be released under the GPL licence.

I think that was adequately covered earlier in the thread, the complications of the individual elements having different licences and varied copyright owners rather rules that out unless you write clones of those elements which are licensed in a compatible fashion. I think that suggestion came from Jeffrey way back up the thread.
One element that comes to mind that needs work but is closed source is Resolver (the whole stack could probably do with work) so if someone wrote a new version (non-GPL because of licence clashes) then I’m sure it would be well received.

So, in a way you’re right; before I was motivated and now I am not.

Oh, that’s 24 hours for you. Hopefully another 24 will mellow things.

Want to run an emulator for a A3000 and play Elite? Well, we can’t help you with ROMs, that’s stealing.

As has been pointed out, that is true. The commercial nature of the ROM image as released, and the licence as included with the product mean that the image is protected in law.
However an emulator for a later system (RPCEmu) is able to run with a freely available OS image and work by a regular contributor round here means more and more of those old games can run in that freely available OS. We aren’t able to change the law relating to those old releases. We are able (subject to talent) to alter new releases. So interestingly we have a situation where it is largely illegal¹ to run an old OS in RPCEmu but perfectly legal to run a brand new OS.

¹ ROL released their “nearly free” soft copy of RO4, but most copies of that are likely to be used with the commercial emulator so the largest body of RPCEmu users with RO4 have probably produced the extract of RO4 ROMs in a technically illegal setup.

I’m persisting because I think it’s a good idea for RISC OS to be free. However if the main-stay of the OS developers (whom I’m presuming are well represented on this forum) disagree, then it’s a bit of a non-starter.

I don’t think people are “disagreeing”. It would be nice if it were “free”. However see below…

“So, here is an idea. Why doesn’t someone at ROOL who knows the person/people behind Castle ask them how much money they want…

As already explained, the IPR isn’t solely and entirely owned by Castle. At minimum there is a residual (and one-time bitter) dispute over who owns what. Involving many parties, some of whom are now out of contact.

So despite the desirability of a free and open RO, the reality would require dealing with a very messy legacy of the way Acorn departed. Again I have to ask if you are really aware of this and all the arguments over many years. You keep simply referring to Castle as if they are the only people we’d need to have agree. If they did, others might also pop up and claim rights with various bases and scopes, etc.

And its not a put-down to invite you to try finding all those involved and getting them to agree. But if you don’t already know, please bear in mind that many long-term users and developers for RO have been though that mill in the past. So may simply prefer to get coding, weary and wary of re-opening old arguments.

Jim

Jim; my point was that GPL provides everyone with the same rights. Yes they are based on and limited by copyright, but when the law is based on the concept of copyright there is no other effective method to do it. The fact that copyright in itself is a very broken system is a whole thread in itself.

When I produce code, which I give others the right to use gratis I want to make sure the rights which I have given are protected and the source (and it’s derivatives) are also freely available under the same terms. This goes back to the 4 primary freedoms which the GPL is designed to protect.

The shared source licence is better than a closed-source licence. It does not provide you with the same protection and rights as a GPL licence though; it’s not even an open-source license by OSI’s accepted definition.

If the OEM terms were removed from the shared-source licence then that would be a good thing. If baby-steps are needed to move RISC OS to open-source then I’d happily support them. Although the fact that you can’t even build RISC OS without purchasing a closed-source compiler is probably a more pressing concern (in the absence of a will to GPL the OS).

Maybe people are jaded by past experience, but the licenses don’t help to attract new blood to the project and that’s a shame IMHO.

Seems that only David B’s post above seems to recognise this.

Not true. I do too. I even propose you a very simple plan to achieve this: pay Castle to kill the OEM license, and RISC OS will be an open source OS.
No need to change the license (you can if you want), no need to change the ownership of the project (you can if you want).

You hate GPL licences; I get it – and I disagree wholeheartedly.

No, he understands GPL:
‘It would be a disaster because alternative licences cannot co-exist with GPL in statically linked code’

The GPL is a good license for the IT market as a whole, but can’t be applied for all (that’s why RMS created the LGPL). Now, I don’t say your project is not good, I say yes, but with a different license. If you want to apply the GPLv3 you’ll need to:

4/ You set up a dual licensing option for people who want to pack closed source components in the ROM file (even worse than the current situation).

Closed source or with other licenses, for example the GPLv2.
So open source, yes, GPL, not really.
Just apply a simple plan

1/ Pay Castle to kill the OEM license. Then the other becomes de facto an open source one (remove all the OEM bits and you’ll see it’s similar to a BSD ‘new gen’ license).

Now, I don’t say this just to say no. I propose you this because I have a better view than you on IP problems.

Ok, now we are getting somewhere. So if I did want to pay Castle to replace their license with LGPL (which is a compromise I could get behind); who do I need to approach? Does any one out there in ‘forum land’ have some contacts?

Also, can anyone think of technical reasons (as opposed to emotional) why LGPL couldn’t be used?

x

Although the fact that you can’t even build RISC OS without purchasing a closed-source compiler is probably a more pressing concern (in the absence of a will to GPL the OS).

Ah, now that item is something that ROOL have taken note of¹ and I think you will find that anyone that offered to do lots of work on modifying the source in such a way that both the legacy compiler and GCC could be used would find themselves getting a pretty warm welcome, not least because the existing compiler option has various limitations as far as modern code development is concerned.

¹ Posts by ROOL staff in these forums have specifically mentioned the idea.

Nota: open source is not a war and Castle not evil. I would prefer them to manage a really open source RISC OS than to put it inside the hands of a newcomer. Of course, to set up a foundation or a consortium would be great. But Castle could make it too.

It’s probably easier to take their work as a start, than to buy the product and start all again. Basically, the personal license without the OEM bits is a NetBSD one. So no need to ask the copyright holders (and there are hundreds of them) for a change of license to go to full open source. Just remove the OEM clause and be a fully certified OSI project.

Problem: you do not even want to consider this option. In fact, you don’t want RISC OS to be an open source project, but a ‘free software under GPLv3’, despite the consequences (viral VS ROM packaging) and the amount of work that would be needed (IMHO, it would be simpler to make a clone of RISC OS around a Linux kernel than to try to convert the RISC OS code to GPLv3).

I repeat: GPLv3 means no possibility to package non GPLv3 things in the ROM file. For Linux, every component is a file. With RISC OS, a file can be a set of components. Second problem: most of the code is statically linked, so you’ll need a LGPL with static exception for all libraries (technically it’s not considered as free software any more). Problem you’ll not be allowed to put these semi LGPL libraries in the ROM file. And in all cases, companies that make embedded products would not be able to pack a binary blob (for example a driver) inside the ROM file. The only solution would be to create a sort of ROMFS. And I’m not sure that it will make the job from a legal point of view.

Also, can anyone think of technical reasons (as opposed to emotional) why LGPL couldn’t be used?

Yes, see my message (above). Best would be BSD or MIT license. EUPL or other things could make the job too, but since Castle license minus OEM is basically a clean verbose BSD license, it seems OK. Viral license is not possible with a project used in embedded systems. Too much problems (see ADSL boxes, see smartphones).

“Oh, he must think everybody gets paid to develop the OS. How cute.” – way to make a statement out of the blue with no grounding. My point was that only musicians would argue that limiting the public’s access to art is a good thing in the same manner than only a professional software engineer would argue that limiting the public’s access to code is a good thing.

Maybe the confusion arose because many of the people you’re arguing with already use Open Source licences for their own software away from the RISC OS source? Rick and I both use the EUPL¹, and I’m fairly sure that Dave H, Colin and Jim use GPL2 for their releases. We presumably already agree with you that it’s a good idea, so you probably need to start pondering why we’re all comfortable with letting RISC OS itself reside under the Shared Source licence that it does.

However if the main-stay of the OS developers (whom I’m presuming are well represented on this forum) disagree, then it’s a bit of a non-starter.

Point of pedantry: I don’t think any of us (not that I’m a mainstay) disagree with you that the concept would be a good idea. Where we differ is that, knowing the history of how RISC OS got to be Shared Source, we can see a lot of hard work and some potentially huge pitfalls on the way to making it all truly Open.

If you’re offering to come on board and do the work (and can make a cast iron guarantee that your advances to the numerous individuals and companies involved won’t lose us any goodwill at all, will not re-open past arguments, and will not prove divisive to an already fractured userbase), then great – show your plan to the ROOL team and see what they say. But if they give you the go-ahead, I don’t see that you’ll be doing any coding for the next ten years: there’s a lot of work to do to make it happen.

Remember that we’re all doing work on RISC OS in our spare time, for free – even the ROOL team themselves. We already know that everyone is stretched very thin, so the general feeling is that there are probably more urgent things to be done at present.

If you go back and read the OP you will see exactly what I was asking:

“So, here is an idea. Why doesn’t someone at ROOL who knows the person/people behind Castle ask them how much money they want for RISC OS under the proviso that, if the licence were transferred to ROOL (and I mean code, trademarks, etc.) it would be released under the GPL licence. We could then start a kickstarter (or use any of a number of fund raising websites) to raise the money and buy RISC OS.”

This might be the other reason that you’re getting a hostile reaction… What do you think happened to get us where we are now? ROOL sat down and talked to Castle, and I’d guess they all looked at the paperwork. Rick’s already pointed you to ROOL’s Shared Source FAQ – the “Why haven’t you used the GPL?” section mentions both the commercial and practical issues that existed at the time (I suspect that it’s the practical ones that remain now, but they’re hardly insignificant by themselves if you’re set on GPL2).

I don’t know how long you’ve been around the RISC OS world, but you’ll forgive me if I tell you that you’re coming across as being unaware of the last fifteen years of RISC OS history (from it being owned by Acorn, to today). Most people here are aware, and realise that a lot of it cramps ROOL’s style somewhat in a lot of areas.

I admit I am presuming that someone lurking around here knows the 3 people who are associated with Castle?

Again, if you knew your RISC OS scene, you’d know the answer to that. At least one of the three is still in touch and providing email support for Castle’s legacy hardware products, and ISTR speaking to him on the ROOL stand at a RISC OS show or two in the past couple of years. I can’t imagine that he and ROOL don’t talk when they need to.

“[…] Oh, you’d like to change the SD Card image? Well, that’s 40 quid – sorry about that, but “excuse here” is the reason we don’t use GCC”

I think you’ll find it’s comments like this that are causing people to post “Wah!” in pink text, because they suggest that you’re making no effort to understand the history of how we got here (which is key to why things are as they are, and cramp people’s style going forward).

The RISC OS source is heavily tied to the Acorn DDE because that’s what Acorn used. Aside from anything else, for at least half of RISC OS’s life at Acorn, GCC and AsAsm (the Open Source compiler and assembler used by the GCCSDK) didn’t exist for the platform: it was Norcroft and ObjASM² or nothing. And if you’ve every used the DDE, you’ll know that it’s difficult to master and horribly non-standard. As a result, a lot of the RISC OS sources would baffle Open Source tools like Make.

There’s also the tiny problem that Acorn worked on RISC OS machines, so the OS build process requires a RISC OS machine to run on. Aside from anything else, it needs a copy of ARM BASIC to be available, to process some of the source files³.

So we’re stuck with the DDE for now, although work is being done to make the sources compile using GCC on RISC OS (you’ll still need BASIC, so you’re still stuck with doing it on a RISC OS box). The ownership of the DDE is probably extremely complex, and making it open source is likely to be just as time consuming as doing it for RISC OS (you’ll probably find that the rights for the Norcroft compiler will be fun to untangle, certainly).

As for charging 40 quid, I’m afraid that I’ve got some sympathy. ROOL need an income to cover the basic costs of the company’s existence, and there’s not much that they can sell – the Bounties don’t seem to bring in much money.

As I say, I think the hostile reaction you’ve received is due to the fact that you’ve appeared (from nowhere?) and demanded that RISC OS be open sourced without listening to people’s answers. If you have some genuine ideas as to how to go about it in a way that doesn’t ruffle feathers and wake up the old arguments that have dogged and destabilised the platform for much of its life since Acorn’s demise, then I think most of us will listen. If you also have serious suggestions for how to make it happen without diverting our already small team away from doing useful things like developing the OS, then even better.

The problem I have with what you’ve written up to now is that you’ve showed no sign that you’re even aware of any of these subtleties – let alone understand them or their significance.

¹ Dunno about Rick, but there’s chunks of my code out there licensed under GPL2, too.

² At three-figure sums back in the 1980s, compared to 40 quid today.

³ About a year ago, I started looking at the possibility of converting all of my old BASIC source code into plain text so that I could store it under version control. If you’re not on a RISC OS machine (which I’m not), it’s “non-trivial”.

Nota: https://www.castle-technology.co.uk/secure/licence/shop/catalog/contact_us.php
Just works :)

Other people heavily invested in RISC OS are 3QD, RComp and CJE.
(sorry, I probably forget a lot of them).

Ok, now we are getting somewhere. So if I did want to pay Castle to replace their license with LGPL (which is a compromise I could get behind); who do I need to approach?

ROOL, please. The last thing that RISC OS needs now is someone with little obvious understanding of the platform’s ‘political history’ diving in feet first and waking up all of the old fractures and divisions.

Does any one out there in ‘forum land’ have some contacts?

This is why I’m sceptical of your ability to do this. If you knew enough about the RISC OS world to be able to do this task ‘safely’, you’d already know the answer to this question.

ETA:

Other people heavily invested in RISC OS are 3QD, RComp and CJE. (sorry, I probably forget a lot of them).

Precisely. Simon: talk to ROOL about this first.

Yes, of course. ROOL was implicit :)

I’m aware of the history behind Acorn and RISC OS (and PACE, ARM, Element14, etc., etc.). I’ve owned Acorn machines since the BBC was first made (actually I have 2 BBCs, an electron, a master compact, an A3000, A 440/1 and two RISC PC 600s). However, I don’t feel constrained by such history – because I’m not and neither are you. That was then and now is now. The commercial landscape of RISC OS is completely different now to even 5 years ago.

My primary area of interest is embedded microcontrollers and electronics (I’m sure you could find my website on the subject if your interested in the specifics). I’ve been busy repairing and restoring old Acorn gear for a while but, when RISC OS popped up on the Raspberry Pi the two paths crossed and my interest level jumped up a notch; usually I just lurch on the Icon Bar forums, but this topic is specifically RISC OS, so I signed up to the forum and posted.

On the subject of build environment, last time I checked (which was this morning whilst compiling some source code) the GCCSDK allows GCC to quite happily run on RISC OS hardware. I have both GCC and the DDE installed on my Raspberry Pi B+. So, in short, I don’t see why we are ‘stuck’ with the DDE at all and I can see elsewhere on the ROOL site others are already discussing and working towards removing it. Ok, so this might be difficult if you want to cross-compile, but even the ability to compile RISC OS on RISC OS hardware (with GCC) would be a step in the right direction. As a side note I don’t see why cross compatibility with the DDE is even a factor. It’s inferior and closed; take a deep breath and let the DDE die, we don’t need it. GCC is open-source, if it lacks something then it can be added.

As for ‘demanding’ RISC OS is open-source; I don’t think that’s a fair view of my intentions. What I presented was an idea to crowd-fund the money required to convince Castle to hand over the rights to ROOL. In return for doing this the ‘crowd’ would get an open-source OS from the deal i.e. ROOL commit that, if the crown-funding is successful and the rights are transferred, then ROOL would release the source under a true open-source licence. If GPL is too difficult to implement then LGPL would be a compromise.

I’m happy to do some leg-work to achieve this. I’m even willing to throw money at it to make it happen (and I honestly believe that there are others in the RISC OS community who would join me).

If I wanted to go boots-first at Castle, why would I have started this forum topic in the first place? I had an idea and I wanted to share it to see if something could be done about it; that’s got to be preferable to sitting back and accepting the status quo?

As a side note I don’t see why cross compatibility with the DDE is even a factor. It’s inferior and closed; take a deep breath and let the DDE die, we don’t need it. GCC is open-source, if it lacks something then it can be added.

First modify the build to make the source compilable with GCC, then ask the obvious question why anyone would want to use a commercial compiler instead of a free one.
While you have a situation where the first item is not complete the follow up question is academic.

There will be a period where both are in use therefore cross compatibility would be required. I’m not sure of current status, but last time I looked the information from the GCC coders was that it couldn’t, yet, do the complete build.
Work in progress and they probably welcome additional help.