ROS must open up to survive says Wild | RISC OS news and information from Drobe


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

ROS must open up to survive says Wild

Published: 19th Jul 2006, 13:53:44.

More to follow later

Castle director and Tematic boss Peter Wild has said RISC OS must become open source to compete in modern embedded markets. Speaking personally, he outlined his views in a comment posted under a drobe.co.uk interview with RISC OS Open shareholder Andrew Hodgkinson.

Pete, pictured right, said: "It is frequently mandated by telcos and service providers in their specifications for STBs [that the operating system] is open source and the telco will have access to a system it can work with and modify independently of the hardware manufacturer.

"In emerging markets without clearly defined standards this is considered essential for future proofing.

"Make [RISC OS] open source and encourage users to contribute to it, use it and drive its ongoing development. This is the only scenario I can see in which the solution might prosper - short of someone investing a few million quid to fund further development."

He added that RISC OS was 'too far behind' to successfully target the mobile phone markets, would be unable to take on WindowsCE in the PDA realm, and would take a hammering from Linux in the IPTV STB markets. This leaves media players and more niche applications - with extensive work being carried out by engineers at Tematic on the media playback front.

Open sourcing RISC OS could spark a growth in development and allow the two streams of the OS to re-merge, according to Pete. He argued this would benefit rivals Castle and RISCOS Ltd., and desktop users.

Pete is also hoping to sell off his shareholding in RISC OS 5 owners Castle, after he loaned them nearly £30,000. He had hoped to build up the company to attract a larger outside form or business angel to swoop in and snap up Castle as an IPR asset - if the operating system were to be made open source, Pete could perhaps make use of the IPR without being tied to Castle.

For all of Pete's musings, scroll down to the comments on this article.

Links
Peter Wild's personal website

Discussion

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

Put the words OpenSource and Risc OS together and all I can think of is Netsurf and Firefox neither of which cover themselves in glory as far as I am concerned. So question 1 would there be enough developers happy to make their contributions for free out there.

Peter raised a lot of issues with his post apart from Open Source making Risc OS free and therefore better able to compete against Linux I do not see how it helps with all the other issues he raised.

I would also question how relevant Risc OS would be in a lot of embedded applications. Routers tend to run Linux because they need multitasking and good TCP/IP support. I would have thought streaming video and music would not lend itself to cooperative tasking especially when you start to think of many TV's and Music outputs in one house.

The fact that he is trying to get away from Castle hardly makes you think that Castle and he see eye to eye about this and Risc OS's future

Jwoody on 19/7/06 4:36PM [ Reply | Permalink | Report ]

"Put the words OpenSource and RISC OS together and all I can think of is Netsurf and Firefox neither of which cover themselves in glory as far as I am concerned."

What an incredibly ignorant and potentially damaging thow-away comment that is. You clearly have absolutely no understanding of the achievement that Netsurf and Firefox (the RISC OS port) represents. The current version of Netsurf demonstrates several years of work by a team of extremely dedicated developers who have received no payment for their work. It's a superb piece of software and brings the RISC OS platform closer than ever to current web standards in an impressively user-friendly and "smooth" package. I'm sure your comments act as a real encouragement to the few remaining developers our platform is fortunate to have.

fylfot on 19/7/06 5:16PM [ Reply | Permalink | Report ]

Jwoody:

Second to my previous post: your comments seem to suggest that the Netsurf project has some how failed. How exactly?

The aim of the project, according to the NS web site, "is to bring the HTML 4 and CSS standards to the RISC OS platform and provide a small, fast and comprehensive web browsing solution."

Well, my goodness, they're almost there! I think the NS project and its developers have every right to "cover themselves in glory" but I think it is our responsibility to thank them for their achievements instead.

fylfot on 19/7/06 5:26PM [ Reply | Permalink | Report ]

netsurf may well have had a lot of work done on it and yes the programmers deserve thanks but it is in NO WAY A COMPETE BROWSER with javascript, flash etc etc. SO SORRY IN MY BOOK THAT'S NO CIGAR.

A open source Risc OS would no doubt take even more effort. Are you saying you want an OS that is only 70% 80% complete ?

Jwoody on 19/7/06 5:35PM [ Reply | Permalink | Report ]

In reply to Jwoody:

... better than a closed OS that is 90% dead, eh?

The momentum of development behind RISC OS (both the OS and third-party apps) seems to have really dropped in the last year or so. Many of us are wasting our energy in petty bickering and fighting rather than in development. How many people have you seen admit to looking lustfully at MacMinis recently? But a fundamental shift as is being suggested could kick-start development, get RISC OS on a variety of new kit, and generally breathe another decade of life into the scene.

And regarding him and Castle not seeing "eye-to-eye" as you put it? Well, all we have here is someone knowledgeable about the RISC OS commercial scene just voicing what many of us have feared. Castle's bread-and-butter work is in embedded systems and set-top boxes. The fact that their boxes run RISC OS is not a factor that their customers particularly need to know - it's hidden. So Castle don't really need RISC OS itself to be pushed forward and developed at the rate that we (the RISC OS advocates) want!

As things stand at the moment, I estimate I'll be getting rid of my RISC OS kit in a year, maybe less. I can't dedicate my time and energy to an OS that may die if the owners decide it doesn't put food on their table anymore. But if the OS was made open-source then I'll know that it will always be around in a useable form, and I'll stay here ... where I'm happy!

krisa on 19/7/06 5:58PM [ Reply | Permalink | Report ]

fylfot: You're absolutely right. In my opinion, NetSurf is currently one of the finest applications to grace the RO platform, even in its pre-1.0 release form.

Jwoody: Cool down man.

NetSurf does not pretend, or was ever meant, to be a 'complete browser with Javascript'. Flash is, as far as I know on about every platform, an external plug-in / application apart from the browser.

"Put the words OpenSource and RISC OS together and all I can think of is Netsurf and Firefox neither of which cover themselves in glory as far as I am concerned. So question 1 would there be enough developers happy to make their contributions for free out there."

Actually, there are a quite a few developers who've released their sources through one or another license and many still do. Further to your 'question 1' - I imagine there are developers from outside the ROS platform who'd find it challenging and interesting to join development if it were to be open sourced.

"Are you saying you want an OS that is only 70% 80% complete ?"

When is an OS, or application for that matter, ever complete? Is your favourite OS complete and how do you know? Perhaps when the design goals have been met and that's certainly the case with RO.

From above and earlier comments I can only conclude you are a troll. Go back to where you came from until you're able to make some constructive comments.

hEgelia on 19/7/06 5:59PM [ Reply | Permalink | Report ]

"From above and earlier comments I can only conclude you are a troll. Go back to where you came from until you're able to make some constructive comments."

In my experience people who go around accusing people of being trolls, are those that don't like the TRUTH. Seems like a standard defence mechinism.

Why should the Risc OS market accept second best. ?

Jwoody on 19/7/06 6:25PM [ Reply | Permalink | Report ]

hEgelia:

"I imagine there are developers from outside the ROS platform who'd find it challenging and interesting to join development if it were to be open sourced."

Also, while porting RISC OS to an alternative hardware platform is an incredibly formidable task, you can bet your bottom dollar that if RISC OS was open sourced people will try - and might eventually (for very long values of eventually) succeed, because we're no longer talking about company resources, finding the funds to pay a team to do the work and having a plan to recoup the cost afterwards. It'll be people doing it off their own back, in their spare time.

VinceH on 19/7/06 6:43PM [ Reply | Permalink | Report ]

"Also, while porting RISC OS to an alternative hardware platform is an incredibly formidable task, you can bet your bottom dollar that if RISC OS was open sourced people will try "

Go look around sourceforge there are loads and loads of open source projects. Some very successful some not. I would contend that the long tail were not. Just making something Open source does not gaurantee success you have to build a good community behind it. Based on expereince so far of Risc OS developers doing open source I would say that as far as the second is concerned the jury is out.

Jwoody on 19/7/06 7:02PM [ Reply | Permalink | Report ]

Jwoody> maybe so, but would you say that commercial development of RISC OS or its applications over the same time period /has/ 'covered itself in glory?' Open sourcing may well be a long way from the ideal scenario, yet still be an improvement.

adrianl on 19/7/06 7:12PM [ Reply | Permalink | Report ]

"maybe so, but would you say that commercial development of RISC OS or its applications over the same time period /has/ 'covered itself in glory?"

Some like Artworks have been good. Risc OS has at least made some progress even if some of the fundamentals have not changed.

Anyway the whole argument is academic unless Castle do something brave.

Jwoody on 19/7/06 7:25PM [ Reply | Permalink | Report ]

For what it's worth:

I am currently back to using RISC OS on my Risc PC. I have installed the RISC OS 4.39 ROMs last weekend and the accompanied CDROM. I am typeing this comment from this RISC PC using NetSurf. Honestly, I work 8 hours a day with a PC at work and believe me I really like RISC OS better.. If it was to be open sourced I for certain would get involved. I might even use that source to get it to run on the PocketPC..

To all people who are annoyed by negative comments I would tell them to cheer up. The world is full of people telling you dayly that ".. can't be done" "... will never happen" "... is impossible" .. There were times when people got burnt at the stake for even believing that the world was round..

RISC OS can have a great future when it is opensourced and Yes there are many programmers to support such an initiative. The important thing to do is to believe you can set a goal and achieve that goal. Find people who are willing to help and ignore those that try to tell you otherwise. People have a huge intellectual potential which often gets wasted by negativity don't let that happen to you.

RISC OS would never run on another platform they said. Emulation has brought RISC OS on Windows, Linux, MacOSX and Windows CE. The next step would be to have a hybrid solution which would need source code to create a native RISC OS compatible layer.

The future is ours. Companies will come and go but good ideas and hard work will create the unimaginable.

Jan Rinze.

JanRinze on 19/7/06 7:40PM [ Reply | Permalink | Report ]

In reply to Jan Rinze:

Well said.

MikeCarter on 19/7/06 8:25PM [ Reply | Permalink | Report ]

If memory serves me, Explan moved their SOLO project to Linux because RISC OS remained a proprietorial product. If RISC OS were to go open source, maybe they would return to the fold. Think of all the potential new users and developers that would bring to the platform.

JWCR on 19/7/06 9:52PM [ Reply | Permalink | Report ]

Jwoody:

"Just making something Open source does not gaurantee success"

Quite right. And you'll notice that in my comment, I didn't say such a thing (which, let's not forget, is only a hypothetical situation anyway) *would* succeed, only that it might - and that possible success only after a considerable amount of time.

VinceH on 19/7/06 10:03PM [ Reply | Permalink | Report ]

Open sourcing the OS would certainly give me faith and reassurance to invest in new hardware. Furthermore, it would make me happier to buy more of the big commercial applications because I will have far greater confidence that the OS will proceed and evolve over the coming years and my investment won't be more secure. I don't know what RISC OS Open are planning, but if it's anything related to producing an Open RISC OS, that would be very amazing indeed.

Ginger2 on 19/7/06 10:48PM [ Reply | Permalink | Report ]

It is true that many open source projects never really take off, because of a lack of developers. However this is usually not the case with open source operating systems. Most open source projects are very specialised programs, so the number of developers interested in the particular application can be relatively low. However operating systems are needed by everyone and cover a wide range of applications. Thus the number of developers that are interested in its development is much bigger to start with. For example Martin Würthner might be interested in improving some OS component wich has to do with graphics handling. This component might interact with some graphics-driver-level code, so he might even improve some of that. I'm sure Steffen Huber would have very much liked to improve the ATAPI driver and ADFS when he tried to get more decent DVD writing speeds. The list goes on.

JGZimmerle on 19/7/06 11:19PM [ Reply | Permalink | Report ]

I agree absolutely. Currently my only real involvement in the RISC OS world is in reading drobe et al in the hope that something positive starts to happen to the platform-I stopped using real hardware years ago and since Windows is not an adequate OS for the sort of work I do, I am not even able to run an emulator. An Open Source RISC OS would lead to myriad possibilities for using the platform in new ways. I believe that there's a substantial number of untapped RISC OS programmers who aren't employed by any of the OS's licensees but who would happily contribute if they could. I admit that I'm ideologically drawn to the FOSS model but from a practical perspective the case has been proved again and again. As an grid computing software engineer the entire suite of software on which I and the entire community develop is open source and though I agree that not all projects succeed in the long term, consideration should be given to what happens when a project dies. In the closed source world when development stops or gets grindingly slow, that's it - finished. With open source software the source code is available for anyone else to give it a go, whenever they feel like it. This really works! The transformation from XFree86 to X.org is a recent example. (Okay, so XFree86 isn't actually dead but development was pathetically slow). If this is even a vague possibility, it is the best news I have heard regarding RISC OS for many many years.

danielhanlon on 19/7/06 11:30PM [ Reply | Permalink | Report ]

I find it amusing that someone at Castle is advocating for open sourcing RISC OS when in the past they forced RO Ltd to withdraw the open sourced version of !Printers. That said l agree with Julian there are a few developers out here with the technical knowledge to seriously improve some parts of OS, especially since it seems that on both Castle and RO Ltd sides manpower is now seriously limited.

andretim on 20/7/06 8:05AM [ Reply | Permalink | Report ]

Castle aren't advocating Open sourcing RISCOS - if they were it would be open sourced as they own the code!

Its Peter Wilds suggestion for the best way forward.

markee174 on 20/7/06 8:11AM [ Reply | Permalink | Report ]

Open source is perhaps the only way to repair the forked OS problem. I'm all for it. Maybe we should opensource some of the older applications as well, especially those whose original developer and/or distributor has left the RISC OS scene. Maybe somebody should take it upon themselves to set up the environment for an opensource RISC OS. I'd be quite happy to see any of my public domain stuff put in there. Surely there could be a considerable number of non castle non ROL modules that can be added as well. This gets the ball rolling. Due to the way RISC OS is organised there is nothing to stop people re-writing modules from scratch.......just leave a big gap between the current commercial version number and the opensource one ;-)

let's start all opensource modules at version 100.

As for the Netsurf criticisms. It is not helpful to deride the work done the way James did on the other hand we should be realistic in that only a browser as capable as for example Firefox (as it works on other platforms) is a realistic proposition to attract new users. RISC OS must be capable of performing all the basic functions that people expect of a computer and let's be honest here : phones and PDA's. If there is one phone or PDA out there that has more computer related functionality than a desktop then that desktop has a serious problem!

One of RISC OS' selling points is the speed of boot up. Given that it has a very small footprint compared to other OS's it has the potential to be the first to take advantage of the emerging fast non-volatile memories e.g. nram, mram, pram. The cost prohibits their use at present but the real cost is the total amount of memory needed NOT the unit chip price. A RISC OS machine that switches on in a second in exactly the same state you switched it off in........that is huge selling point. With some clever shuffling of the way RISC OS handles memory couldn't the "boot up" state be in the MRAM and the much cheaper DRAM used for dynamic areas....for example. Pedantics please note I know the details of teh OS at present preclude this but you get the point I'm making.

mripley on 20/7/06 8:53AM [ Reply | Permalink | Report ]

I too can see the point that 'Jwoody' is making about NetSurf. Look at it from a generic user's point to view (i.e someone with no experience of RISC OS). They expect a browser on a desktop computer to display every single page they access without problem. If it doesn't then it's not a good browser. Furthermore, they don't care about the internal structure (core features vs. plugins etc.), or how much effort went into developing it. Whilst NetSurf may be a big achievement in the niche RISC OS world, Peter Wild's original statement was talking about RISC OS in a wider context.

dms on 20/7/06 9:37AM [ Reply | Permalink | Report ]

Open-sourcing RISC OS seems like a very good idea. Not all forms of open source preclude licensing, so RISC OS Ltd. and Castle could still require license payment from users, and if the source is open for inspection and modification, RISC OS would be more acceptable to OEMs, and you would see many more independent people making patches and bugfixes, which can be integrated into the core OS. It could certainly potentially speed up movement of RISC OS to new ARM-based platforms, such as PDAs or handheld games consoles (like the GP2X).

You would need to find a license form that allow the owners to retain rights even if they later include contributions from third parties into the core OS, though. Or release the code into public domain, but I can't really see this happening in the near future.

torbenm on 20/7/06 9:46AM [ Reply | Permalink | Report ]

Many commercial companies release code under a dual GPL or commercial basis and only accpet patches if they can include the code in commercial versions. MySQL is the classic example.

markee174 on 20/7/06 9:50AM [ Reply | Permalink | Report ]

I use a variety of browsers on RISC OS for different sites. Netsurf is not yet the finished product, but for most sites that don't use JavaScript, it's the best RISC OS browser (in my opinion). I would hate to be without it and hope its developers haven't read some of the negative comments on this thread and been tempted to pack it all in.

They are giving their free time to produce a quality browser for our use. They said all along that JavaScript support was a low priority. Maybe one day they'll be able to bring it to the top of their priority list, but for now let's just thank them for giving us the best browser our platform has got.

cables on 20/7/06 12:26PM [ Reply | Permalink | Report ]

Whether Firefox/Netsurf prove (or disprove) the usefulness of "open source" is neither here nor there.

As to the substantial point of open sourcing the OS is that *really* a runner? The number of developers here in RISC OSland is somewhat limited - so the scope for large "community based" initiatives are limited. There is also the question of co-ordination (if we can't get a coherent approach with just *two* variants of RISC OS what of 3, 4 or 5 versions or subversions as different developers find new ways to produce a "re-invented wheel with unique features").

You'd have some slow OS improvement - and absolutely or almost no Application development (and who wants an OS without a decent WP/Browser/<insert your desired application type here>

.

As to the argument that telcos require an "open source OS" is - if you don't mind me saying it - misleading. You'd probably find that they'd be happy enough with an OS that adheres to "open file formats/open comms protocol use" and where the source is kept in escrow so that (if) the OS vendor goes belly up that the telcos aren't left twisting in the wind.

As to the argument about getting RO running on different platforms most of these are *not* manufactured by Castle/Ad6 or anyone with a *real link* with the RISC OS community. Many come with PalmOS or WindowsCE (which you pay for even if you *do* replace them with ROS) and that means that the limited hardware spend people have will be "spread out" to encompass companies that have *no* need/wish to support ROS users. It would effectively *kill* the desktop RO machine market - I'd consider that a very bad thing.

RO5 already is sufficiently hardware abstracted that it could support newer ARM chipsets - it would be best to stick with and develope that rather than dissapate so much energy advocating an approach that would (IMHO) lead to further fragmentation, waste of effort and with the potential loss of the desktop RISC OS hardware sector.

AMS on 20/7/06 1:19PM [ Reply | Permalink | Report ]

andretim: "[Castle] forced RO Ltd to withdraw the open sourced version of !Printers."

Except it was never done -- source still available at [link]

ben on 20/7/06 1:23PM [ Reply | Permalink | Report ]

I'm sorry that this isn't strictly on topic, but I think !Netsurf is great. The folks who develop it deserve a lot of credit. I tried downloaded Firefox once, but my olde ARM710 is too slow to use it, but it was an excellent job getting it to run on RISC OS.

My RISC OS fantasy would be RISC OS X, something along the lines of MacOS X, but with RISC OS.

acornnutcase on 20/7/06 1:40PM [ Reply | Permalink | Report ]

And there lies the problem. My old ARM 710. Get a faster machine. :-D

Cheers Bob; who found the SA riscpc too slow.

nijinsky on 20/7/06 2:16PM [ Reply | Permalink | Report ]

I think open source is a good idea /if/ there is someone to manage the project. Otherwise it becomes a hotchpotch of people working on bits that interest them rather than co-ordinated development on key areas that need work. The current situation with ROL and Castle driving things plus open source RISC OS would work, if (big if) they can come up with a funding stream.

I'm amused, though, that ROL are selling RISC OS specially for its closed-source-ness: [link] says:

> Proprietary code > RISC OS is based almost entirely on proprietary > code with no chance of hidden backdoors that > can arise from using Open Source software.

caliston2 on 20/7/06 6:11PM [ Reply | Permalink | Report ]

That statement in ROL's brochure shows they either have no idea at all about open source, or are desperately splashing around. Or most likely, both.

moss on 20/07/06 6:48PM [ Reply | Permalink | Report ]

AMS: "RO5 already is sufficiently hardware abstracted that it could support newer ARM chipsets - it would be best to stick with and develope that rather than dissapate so much energy advocating an approach that would (IMHO) lead to further fragmentation, waste of effort and with the potential loss of the desktop RISC OS hardware sector."

I guess you didn't read the article, then, or didn't get what Mr Wild was trying to say: in order to really make his investment worthwhile, RISC OS has to look beyond a few flavours of ARM-based equipment. Of course, Castle and friends can keep slogging away in a part of the business whose size relative to the total market is probably diminishing, with an operating system that doesn't cover all the hardware options for customer projects, thus ruling it out as a common platform (which is where Linux comes in, by the way) where the customer and their partners can bring the same (or related) expertise to bear in all projects throughout all the different niches and sectors in which they operate.

Want to do a cell phone using an ARM chipset and a GPS unit using a MIPS chipset? Need RISC OS and something else for the MIPS stuff? The ARM chipsets don't cut it for the GPS stuff? Perhaps we can find something else that actually does both ARM and MIPS.

If RISC OS were open source, you'd firstly get some pretty qualified people looking at it, some of them perhaps re-motivated to give it a second look. And you can hardly criticise open source project management techniques in the light of the way the RISC OS scene has been going over the past few years: the captain and the admiral arm-wrestling for a meagre haul of treasure on the main deck while the ship's wheel turns all by itself.

guestx on 20/07/06 8:20PM [ Reply | Permalink | Report ]

GuestX>Actually I was quite impressed by Peter Wild's contribution - however I wouldn't necessarily agree that the solution is simply GPLing the whole kit and kaboodle.

Amongst other things he wrote "RISC OS’s biggest plus is also its Achilles heel - it only runs on ARM (obviously). With 70% of the OS kernel being hand optimised ARM assembler it has an enormous performance benefit in its core functionality over compiled OSs such as Linux, VXWorks or Nucleos"

We *lose* big plus the moment we code in a high level language so that the applications and OS can be run on a processor other than ARM. It probably also isn't really necessary as later ARM's (such as the Cortex) may well offer significant speed boosts.

While I respect Linux developers their interest is *Linux* not *RISC OS* so why should they chip into help ? That leaves the RISC OS developers - but are there enough to make this viable ?

If RISC OS was opened sourced (in the GPL sense) what you'd probably find is any useful bits of RISC OS would be "converted" and dropped into Linux - and all that would happen is you'd have a more enhanced Linux taking more of the embedded/handheld device share of the market in competition with "open" RISC OS which would slowly die.

If you asked me what *could* be done. Well I'd say that the developers of the OS (e.g., Castle/ROL) should provide a stimulus for development by identifying areas where the OS could be enhanced - and granting experienced developers access to source for specific upgrades. If the changes are accepted the OS vendor "rewards" to the developer (money+acknowledgement/discounts on the vendors products etc.,). This work could be organised in a collaborative manner by small groups with team leaders appointed by the vendor (in effect this leverages the developers they have - acchieving more than they could with the current limited numbers of internal staff). Another approach would be to launch "competitions" for developers to produce OS enhancements (again with access provided to source where required under conditions of non-disclosure).

This sort of approach may take some of the better aspects of the "open source" style of development - without endangering the future of RISC OS.

AMS on 20/07/06 9:27PM [ Reply | Permalink | Report ]

I think Peter Wild makes a well-reasoned case, but one thing puzzles me somewhat: since on his own admission he has been closely associated with Castle and Tematic for a number of years, what was his view while Castle/Tematic were vigorously pursuing the proprietorial OS route, i.e., trying to establish sole ownership of and license to RISCOS? Did he support this strategy at the time? If so, why has he changed his mind?

bucksboy on 20/07/06 10:55PM [ Reply | Permalink | Report ]

AMS> The speed loss you anticipate from converting to a high level language is far from obvious to me. In fact, in a high level language, programmers are more willing to use improved algorithms and to improve the existing code. Furthermore, by running the code through a new compiler, all code can be tuned for a new processor pipeline (particularly important on the XScale, and even StrongARM). This is something that still hasn't - and almost certainly never will - be done for much of RO's assembler code because it would be a lot of work.

Whilst there may not be enough developers out there to make dramatic improvements to RISC OS, were it ever open sourced, they are still a superset of the developers that have tried to work on it commercially over the past few years, many of whom have been forced to depart.

adrianl on 20/07/06 11:23PM [ Reply | Permalink | Report ]

@AMS: Sure, some Linux developers are only interested in Linux, but most of them (at least most of the ones I know) are very open-minded people, who just like the idea of advanced open source software. I personally know three Linux developers, who are very interested in RISC OS. Two of them have already bought second hand RISC OS machines from me. As for the use of RISC OS source code in Linux, I'd say this would be a good thing. Why should anyone re-envent the wheel? In the past be have profited greatly from other open source projects, so why not return the favour? In turn we could make use of a lot more open source code, because suddenly we would not violate the GPL anymore. Your suggestion of a development model looks to me exactly like the one RISCOS Ltd use. And it kills the feel-good factor of developing something for the good of man-kind, because ordinary users still have to pay for the software, instead of paying for service and support.

JGZimmerle on 20/07/06 11:24PM [ Reply | Permalink | Report ]

Without wanting to get embroiled in the open source debate, I'd just like to point out that if you have the source to the entire OS and the limited application set to run on a specific device (rather than everything thats ever been written for the RISC OS desktop), you can port to other processors without having to re-write large portions of ARM assember in a high level language. There is no reason why ARM assembler cannot be cross compiled to different architectures, its a very small clean ISA which can be treated almost like a byte code. ARMs strengths such as the 'free' barrel shifter and conditional instructions can be used as hints to extract parallelism and fill branch delay slots in other architectures.

druck on 21/07/06 09:37AM [ Reply | Permalink | Report ]

Iv just been thinking, that if RISC OS was ported to other processors such as an x86, it would no longer be able to be called 'Reduced Instruction Set' Computer OS, cause it wouldnt be runing a a RISC processor. please correct me if im wrong.

MikeCarter on 21/07/06 10:42AM [ Reply | Permalink | Report ]

Mike: I find it amusing that those who have been critical of VA now seem to be positive about an Open RISC OS that would inevitably dump ARM processors. To me RISC OS on ARM is important for all the well-rehearsed reasons and my acceptance of VA is predicated on the fact that it's second best. You have noticed that the Emperor's New Clothes have a contradictory word set - but that's the least of the problems. Perhaps someone should take a look at the new version of Windows to see what Microsoft have identified as one of the big problems of the future (this week's New Scientist may add a clue). It's possible that RISC OS could play an important role in the future - but that's less likely if it's taken over by those intent only on getting it to run on 'standard' hardware.

jc on 21/07/06 11:37AM [ Reply | Permalink | Report ]

In reply to John:

"It's possible that RISC OS could play an important role in the future"

I also believe this too. Look at your average Joe all he wants a computer for is word processing, data input into spreadsheets, listening to music, do some graphical work. RISC OS does all of these. Also it has been mentioned many times before, that if a new flash player was written for RISC OS, then there could be a big future in interactive displays in museums, college/work reception areas.

MikeCarter on 21/07/06 11:50AM [ Reply | Permalink | Report ]

oh I forgot to mention after the average Joe thing that he an do all those things at nice fast pace, in confidence that half way thought writing an important document the word processor will not run out of memory, crash and loose all his work.

MikeCarter on 21/07/06 11:53AM [ Reply | Permalink | Report ]

John: "I find it amusing that those who have been critical of VA now seem to be positive about an Open RISC OS that would inevitably dump ARM processors."

Well, at least some of that critisism comes from the fact VRPC still only runs on Windows and due to its particular copy-protection. Regarding an open source RISC OS, thus far it has been purely speculation. So I'd say it's certainly not 'inevitable' that ROS would switch to another processor architecture. However, it should be noted that ARM compatible CPU's have typically become unsuitable for desktop computers, even if ROS is largely written for its architecture. Perhaps dual-core ARM CPU's and the forthcoming Cortex series could help compensate for its lack of raw processing power.

"It's possible that RISC OS could play an important role in the future - but that's less likely if it's taken over by those intent only on getting it to run on 'standard' hardware."

Why exactly? It has become clear in the 1990's that this 'standard' hardware would be most succesful, because practically any manufacturer can develop its hardware, ie. it is an 'open' standard, coupled with the dominant use of Windows. For that reason, many 'alternative' OS's have eventually been written or ported for use with this standard hardware, lately even Apple computers. That's ofcourse also a big reason why Linux has become so popular. If ROS were to run natively on x86 standard hardware, it would mean almost anybody could run it on their (cheap) machine, but it would also be 'one out of many' thus particularly dependent upon the merits of its GUI / applications.

Mike: "Look at your average Joe all he wants a computer for is word processing, data input into spreadsheets, listening to music, do some graphical work."

Well, yes, several years ago. I'm sorry, but many people in the market for a computer have a pretty good understanding of what's possible with these things nowadays. People like to burn films now, use all audio / video codecs and data formats thrown at them, have nice 3D accelerated graphics, etc. Just having a modern Flash player on ROS is not going to solve its fundemental problems, namely its severely dated architecture coupled with a deminishing user- and developer base. RISC OS is still a co-operative multitasking microcomputer operating system, it's just not up to the tasks expected by modern day computing, which has in some ways little to do with the processor it runs on.

I believe we can't expect Castle and ROLtd to modernize it to the standards required today. They are too small with too few developers working on it. Castle's interests with ROS lie in the embedded sector, not particularly in the desktop area. If Castle can adopt one of its developments for the embedded / STB / industrial sector for desktop use, do we get to see progress where we like it. I'm not sure about ROLtd, but I believe they are even less capable than Castle. Open sourcing it may be the only chance it's got to revive and become relatively equal in terms of capabilities compared to the likes of Linux, Mac OS X or Windows, never mind a chance in the mainstream.

"oh I forgot to mention after the average Joe thing that he an do all those things at nice fast pace, in confidence that half way thought writing an important document the word processor will not run out of memory, crash and loose all his work."

That nice fast pace and perceived stability can become quickly compromised once one badly behaved application takes over the system which, although reduced in possibility due to the enhanced memory protection in recent versions of the OS, still remains a distinct possibility due to ROS's cooperative multitasking. I've not had something like that happen to me in Linux, OS X or even recent versions of Windows, while similar did happen to me a couple of times in RO4; one naughty application crashed dragging the whole system with it, including my carefully designed, but unsaved, sprites. Until this, and other limitations, are resolved, there's little chance of a 'big future' in the desktop computing domain.

hEgelia on 21/07/06 1:35PM [ Reply | Permalink | Report ]

In reply to hEgelia: "Well, yes, several years ago. I'm sorry, but many people in the market for a computer have a pretty good understanding of what's possible with these things nowadays. People like to burn films now, use all audio / video codecs and data formats thrown at them, have nice 3D accelerated graphics, etc." I feel this case is often overstated. Possibly Many people would like to view DVDs, true, but until Cino appears, that can be done just as well with a stand-alone DVD player that people would probably own anyway. A few people would want to view digital video, but if/when Cineroma sees the light of day I doubt the video formats it fails to support will cause headaches for anybody who can't view them. I don't really see many people burning Video to DVDs or even that many playing games (if they like games, most people seem to have a games console or three). The majority of the time the majority of people seem to use their PC for web browsing and email (the latter probably web-based), downloading and playing mp3s, viewing/uploading pictures from their camera and maybe a little bit of word-processing. Nothing too demanding for RISC OS. Given a full-featured browser, Flash and maybe Cineroma I doubt anyone would really miss anything (though an Excel-compatible spreadsheet and a better Adobe Acrobat reader wouldn't go amiss IMO).

"That nice fast pace and perceived stability can become quickly compromised once one badly behaved application takes over the system which, although reduced in possibility due to the enhanced memory protection in recent versions of the OS, still remains a distinct possibility due to ROS's cooperative multitasking. I've not had something like that happen to me in Linux, OS X or even recent versions of Windows, while similar did happen to me a couple of times in RO4; one naughty application crashed dragging the whole system with it, including my carefully designed, but unsaved, sprites." I'm sorry, but if alt-break didn't work, pre-emptive multitasking would not have solved the problem. Neither does cooperative multitasking prevent Realaudio or 3D accelerated games or DivX decoding on RISC OS.

Cogs on 21/07/06 2:53PM [ Reply | Permalink | Report ]

In reply to the previous post.

We do not have a full-featured browser, Flash or a video player (Cineroma) and we do not seem to any nearer getting them than we were 1 year ago. Projects like Cineroma could/should be sponsored by ROLtd and made part of the core OS and delivered as part of the next Select release, if we ever get one.

Pete on 21/07/06 3:59PM [ Reply | Permalink | Report ]

Cogs: "I feel this case is often overstated. Possibly Many people would like to view DVDs, true, but until Cino appears, that can be done just as well with a stand-alone DVD player that people would probably own anyway."

People expect a new computer to be able to play and burn DVD's where I come from (which is the Netherlands, so that may explain it). Admittedly, there may be a few people around not particularly expecting that, but they are few and far between. Remember, we are talking about people buying a new computer and potentially interested in an Iyonix or A9home. Cino has been in the pipeline for how long now? At least 2 years or so? I remember a Drobe article mentioning an effort to increase data transfer through ADFS, which would help a lot:

[link]

I'm not sure, but I believe that work has not yet been done. There's also the question of money - CinoDVD vs stand alone player. However, I like to mention that there are a lot of instances imaginable were one would like to be able to play a DVD on one's computer. One example; I have a DVD player which I'll give away once we have replaced this RiscPC with a Mac (which, like Linux and Windows, comes with player software).

"A few people would want to view digital video, but if/when Cineroma sees the light of day I doubt the video formats it fails to support will cause headaches for anybody who can't view them."

A few people? Well, again, in my surroundings people actually do want to be able to play or create video's. Cineroma would indeed be very nice, but does it support QuickTime for example? Or other closed source, proprietary formats which are typically found littered over the web or p2p networks? How about video streaming? I haven't heard anything about it for a long time, let alone a recent demo of its capabilities. I suggest you take a read through its FAQ:

[link]

"I don't really see many people burning Video to DVDs or even that many playing games (if they like games, most people seem to have a games console or three)."

Indeed, hardcore gamers tend to either have a beefy Windows box or a game console or three. Still, some of us would like to play a nice game for diversion once in a while. Anyway, we have Doom, Quake and some lovely others like Crazeeman.

"The majority of the time the majority of people seem to use their PC for web browsing and email (the latter probably web-based), downloading and playing mp3s, viewing/uploading pictures from their camera and maybe a little bit of word-processing. Nothing too demanding for RISC OS. Given a full-featured browser, Flash and maybe Cineroma I doubt anyone would really miss anything (though an Excel-compatible spreadsheet and a better Adobe Acrobat reader wouldn't go amiss IMO)."

Many people indeed do. Perhaps it has to do with my age, the particular area's I work and communicate in, but the things you mentioned above are actually the bare minimum. We like to create movies, burn them, professionally process images, audio/videochat, stream live broadcasts, 3D modelling, run several softsynths, convolution reverbs, record multi-channel audio, ..well those last bits are generally music production oriented, but the ones before actually reflect what's considered fairly normal. Just approach a student and ask what (s)he and her / his collegues do with their computers except the usual browsing, chatting, music playing. From all the people I know, I have the most interesting, though least capable computer!

In the end, I think it comes to what people are used to in their surroundings. If I take a look at my father and his contemporaries, what you describe is exactly what he tends to do with his computer. If I look at people one generation younger, they require their machines to do a lot more, like Skyping with their relatives or burning DVD's. I do not wish to imply that you're old or anything, simply that times are changing rapidly in the computer world and ROS has fallen behind quite a bit. When I take a look at pictures of ROS shows, I see many people in their fifties or older, which might explain some things.

"I'm sorry, but if alt-break didn't work, pre-emptive multitasking would not have solved the problem. Neither does cooperative multitasking prevent Realaudio or 3D accelerated games or DivX decoding on RISC OS."

I'm not so sure about that first part, but generally speaking RISC OS itself is dated, apart from its great GUI. I should make it clear that ROS seldomly crashes on me by the way, though I usually do harmless things with it. I did not, nor ever have, implied that ROS' multitasking approach prevents RealAudio, 3D accelerated gaming or DivX decoding from working. I've understood for certain tasks CMT is even preferable, but I think for general desktop use PMT is the way forward.

hEgelia on 21/07/06 4:19PM [ Reply | Permalink | Report ]

Yes we need video support I would like to view things like this [link] With my RiscPC/Iyonix but I can not. I also would like to watch the video's I take on my camera and maybe edit them a bit but I can not. However this is achieved I will be grateful

Pete on 21/07/06 4:39PM [ Reply | Permalink | Report ]

Druck "There is no reason why ARM assembler cannot be cross compiled to different architectures"

Why not start an Open source project to create a cross compiler for ARM.

At the very least it would encourage the powers at be to Open Source. Even if not they could cross compile for a proprietry solutions.

Jwoody on 21/07/06 5:13PM [ Reply | Permalink | Report ]

Druck "There is no reason why ARM assembler cannot be cross compiled to different architectures"

Why not start an Open Source Project to create a cross complier!!!

Jwoody on 21/07/06 5:17PM [ Reply | Permalink | Report ]

John Cartmell wrote>"I find it amusing that those who have been critical of VA now seem to be positive about an Open RISC OS that would inevitably dump ARM processors. "

No there is nothing inevitable about "open sourcing" RISC OS causing it to be "ported" to other processors (although having it open sourced would *increase* the chance of that). David Ruck was positing how hypothetically ARM code *could* be re-coded for another architecture. Given his creation of the excellent Armalyser I think he'd be ideally positioned to comment on the theoretical possibilities of such a venture. I don't believe he was suggesting that it *should* be done but rather how it *might* be done.

JC also wrote>"It's possible that RISC OS could play an important role in the future - but that's less likely if it's taken over by those intent only on getting it to run on 'standard' hardware. "

That is *RICH* coming from you..... You coined the phrase "Hybrid computer" and waxed quite lyrical about VA running on machines that are *precisely that* 'standard hardward'. So remind me again how is that different from what the others are suggesting...... I'd also point out that the processors suggested by others *did not* only specify x86 - Peter Wild (for example) suggested Mips - an honourable and well known RISC processor.

I'd also caution you that VA means (in effect) embracing MS-Windows and becoming dependant on Microsoft - porting to non-x86 or even x86 (sans windows) would *not* involve that. That having been said I can't find much enthusiasm for the notion of RISC OS being moved off ARM - it's what it runs best on and is where (IMHO) it should remain.

AMS on 21/07/06 7:24PM [ Reply | Permalink | Report ]

adrianl wrote>"The speed loss you anticipate from converting to a high level language is far from obvious to me."

There surely would be some, if (say) 1 in 5 instructions were branches to call support functions in the compiled code - if there wasn't much "inlining" and so forth surely there'd be some hit. After all why would anyone *willingly* welcome the suffering of coding in Assembler when a *high level* language with equivalent performance is available.

I accept your point about instruction ordering and compiler optimisations for specific pipeline architectures. And given newer ARM's such as Cortex this is likely to be even more important in future.

adrianl wrote>"Whilst there may not be enough developers out there to make dramatic improvements to RISC OS, were it ever open sourced, they are still a superset of the developers that have tried to work on it commercially over the past few years, many of whom have been forced to depart."

That also I would agree with.

AMS on 21/07/06 7:35PM [ Reply | Permalink | Report ]

In reply to hEgelia: "People expect a new computer to be able to play and burn DVD's where I come from (which is the Netherlands, so that may explain it). Admittedly, there may be a few people around not particularly expecting that, but they are few and far between." Among the general population or just geeks? I don't think I know of anyone burning a DVD with video. Not a single person of any age. They could have and I just haven't heard about it, but it can't be common. When I go out I don't see huge numbers of people with digital camcorders at events. Most people would only produce very mundane video anyway, so I'm not suprised noone seems to bother. As for ripping commercial copy-protected DVDs: that would require software that non-geeky types wouldn't usually have or know where to obtain. I see lots of people with ripped CDs, but not DVDs.

"Well, again, in my surroundings people actually do want to be able to play or create video's. Cineroma would indeed be very nice, but does it support QuickTime for example? Or other closed source, proprietary formats which are typically found littered over the web or p2p networks? How about video streaming? I haven't heard anything about it for a long time, let alone a recent demo of its capabilities. I suggest you take a read through its FAQ: " I don't know much about cineroma, but DivX is about all I require. I steer well clear of Real and WMVs (I think I've caught WMVs trying to do dodgy things and Realplayer wanted my personal details, so I haven't got it - I find I can live without either very easily). As for Quicktime, the cineroma supported codecs list mentions SVQ3, 3ivX, MPEG4 and CVid which I believe covers just about everything you're likely to find in a MOV file.

"Perhaps it has to do with my age, the particular area's I work and communicate in, but the things you mentioned above are actually the bare minimum. We like to create movies, burn them, professionally process images, audio/videochat, stream live broadcasts, 3D modelling, run several softsynths, convolution reverbs, record multi-channel audio, ..well those last bits are generally music production oriented, but the ones before actually reflect what's considered fairly normal." I think the problem is that geeky people tend to have geeky friends. One tends to assume oneself and ones friends are the whole world. I briefly worked for a well-known electronics retailer not that long ago and I know the technical prowes of the general public. The things you list above would be completely beyond the majority of the population. I don't think even most geeky people do 3D modelling - something of a niche requirement there, surely.

" Just approach a student and ask what (s)he and her / his collegues do with their computers except the usual browsing, chatting, music playing." I know lots of students. If they're doing comp sci or digital graphics or something audio or video related, you might be right, especially if they are male. Most students are not those students though. Ask a language student or a sociology student or a history student or a maths student or even an engineering student and you will more likely get the answer I gave.

" From all the people I know, I have the most interesting, though least capable computer!

In the end, I think it comes to what people are used to in their surroundings. If I take a look at my father and his contemporaries, what you describe is exactly what he tends to do with his computer." Funny you should say that as my father is in his 60s and produces a lot of graphics on his computer for a living... In Artworks on a Risc PC. Still, I put a PC together for him a couple of years ago, primarily for Word file import and web browsing. Curiously the two main things my girlfriend does on her PC and she's 25. These are the people we should be able to promote RISC OS to, not the power-users. For that we need a 100% capable browser first and foremost.

" If I look at people one generation younger, they require their machines to do a lot more, like Skyping with their relatives or burning DVD's."

Skype's a good example of something that turned out to be a bit pointless for me. I can call Germany or Hungary for 1p/min from my normal phone by dialling a prefix. Why would I want to pay a similar amount and have to turn on a computer to do the same thing? Local calls are free anyway and I can wander around the house with my existing cordless handset.

"I do not wish to imply that you're old or anything, simply that times are changing rapidly in the computer world and ROS has fallen behind quite a bit. When I take a look at pictures of ROS shows, I see many people in their fifties or older, which might explain some things." I don't see why that matters to the survival of RISC OS. It's actually a potentially larger and more affluent market demographic.

As for preemptive multitasking: it doesn't prevent lockups. Fatal lockups happen when something scribbles all over some unprotected memory in module area or when module code gets stuck in a loop or loses track of its stack due to a bug. I'm sure someone will step in with a correction to the rule, but generally I would say: if you can't alt-break out of it, you wouldn't be able to preempt out of it either. But you generally wouldn't be able to preempt existing apps anyway which is the clincher.

Cogs on 21/07/06 10:37PM [ Reply | Permalink | Report ]

In reply to Peter Darnell: "Yes we need video support I would like to view things like this [link] With my RiscPC/Iyonix but I can not." It actually uses Flash, so what we need is a version of Flash which supports video codecs. A RISC OS video player on its own wouldn't be enough.

"I also would like to watch the video's I take on my camera and maybe edit them a bit but I can not. However this is achieved I will be grateful" Video editing on RISC OS seems to be dead. Whatever happened to Empire?

Cogs on 21/07/06 11:08PM [ Reply | Permalink | Report ]

Cogs: "Among the general population or just geeks? I don't think I know of anyone burning a DVD with video. Not a single person of any age. They could have and I just haven't heard about it, but it can't be common."

That's nice, I guess there's a great concentration of geeks around here then. The Dutch government named our city a 'brainport', because we're quite technologically oriented - it's the city where "Philips" is based, Eindhoven. There's really a lot going on here, so that may explain it. I'm not saying everybody does it, only that it's not hard to find someone putting a recorded video on DVD.

"When I go out I don't see huge numbers of people with digital camcorders at events. Most people would only produce very mundane video anyway, so I'm not suprised noone seems to bother. As for ripping commercial copy-protected DVDs: that would require software that non-geeky types wouldn't usually have or know where to obtain. I see lots of people with ripped CDs, but not DVDs."

Neither do I see "huge numbers of people" doing that, but I do see it regularly at events. As for ripping commercial CD's and DVD's, it (sadly) happens around here judging from the copied discs I notice. In fact, I went out yesterday and I got a DVD from somebody, no rip by the way, just a production. It's really a common occurence around here, but I'm sure if I visit a village several kilometers out of town these things would be unusual.

"I don't know much about cineroma, but DivX is about all I require. I steer well clear of Real and WMVs (I think I've caught WMVs trying to do dodgy things and Realplayer wanted my personal details, so I haven't got it - I find I can live without either very easily). As for Quicktime, the cineroma supported codecs list mentions SVQ3, 3ivX, MPEG4 and CVid which I believe covers just about everything you're likely to find in a MOV file."

I just want file compatibility, that's all. There's a lot of Mac use around here, so QuickTime compatibility is a must. I find this is a nice player for several platforms, which could someday be ported to ROS perhaps - [link]

"I think the problem is that geeky people tend to have geeky friends. One tends to assume oneself and ones friends are the whole world."

That first part may well be, but the second sentence is overstating it in my opinion. Ofcourse it heavily influences my perception, but certainly not to that degree! You know, when I think about it, I suppose I have geek properties, but so have so many around here - strange really... We have a lot of cultural events here, a lot of festivals especially in the summer period and dance parties. Everywhere on these events one sees people young and old working with computers to achieve these 'high end' uses. I think many don't know a lot about computer history and hardware, but they definitely know what can be achieved with them.

"I briefly worked for a well-known electronics retailer not that long ago and I know the technical prowes of the general public. The things you list above would be completely beyond the majority of the population. I don't think even most geeky people do 3D modelling - something of a niche requirement there, surely."

3D modelling and music production uses are more uncommon, but I believe in this area generally speaking people are quite well educated in modern uses of computers. There's even one part of the city completely dedicated to Art in all kinds of forms, including digital means. Yesterday night I sat in a bar where, next to the usual DJ, there sat some people projecting snippets of animation, images, graphics live. Now this is ofcourse not where everyone in this city hangs out, but it's a normal occurence. I really believe that generally speaking, the people in this city use their machines for a greater variety of purposes then ROS currently allows. Perhaps this really is Geek City.

"I know lots of students. If they're doing comp sci or digital graphics or something audio or video related, you might be right, especially if they are male."

Well, next to Technical University we have a Design Academy and other institutes which typically involve using computers to play video, design, videochat across buildings, whatever. We actually have videophones in town, so I'm sure we're mad.

"Most students are not those students though. Ask a language student or a sociology student or a history student or a maths student or even an engineering student and you will more likely get the answer I gave."

Perhaps where you're from. You know, I won't pretend to know what everyone does with every computer around here, or the universe, but when I see a granny with a laptop in the park, what can I say? I must take comfort with the fact that surely someone, somewhere in town uses his computer only and exclusively for 'mundane' purposes like just type a friggin' letter and visit MSN.com. However, until now that seems a rarity, sorry.

"Funny you should say that as my father is in his 60s and produces a lot of graphics on his computer for a living... In Artworks on a RiscPC."

Sounds great. There's a well known guy in the dutch ROS scene called Henk Huinen who also makes (or made..) his living with ArtWorks. Terrific app.

"Still, I put a PC together for him a couple of years ago, primarily for Word file import and web browsing. Curiously the two main things my girlfriend does on her PC and she's 25."

OK, that's nice. My girlfriend is 27 and she likes to play Doom on this RiscPC, next to the usual NetSurf / Grapevine use. Perhaps curiously, but she's been the one complaining the hardest about getting a new machine to be able to process photos, play Doom III, create a film or something. Things most of which, sorry to say, is not yet possible on an Iyonix or A9home. So, we are getting a Mac, since they're nice and easy to get around here. I'll keep a RiscPC in the house though

"These are the people we should be able to promote RISC OS to, not the power-users. For that we need a 100% capable browser first and foremost."

Yes, for now but remember standards are really being raised. Generally speaking you could very well be able to suggest an Iyonix or A9home to a relative, but I tried and it was turned down rather quick. By the way, that was with my girlfriends' parents, not a design student or something. They simply want to do 'normal' stuff, like being able to play something they come across on the web, burn discs, anything - I'm not saying they'll be Skyping all day long or something, just be able to do these things which are not necessarily just "web browsing and email (the latter probably web-based), downloading and playing mp3s, viewing/uploading pictures from their camera and maybe a little bit of word-processing." Furthermore, they don't want to have to shell out for apps which come standard on all the other computers and by that I don't mean the few examples mentioned above, which can be done pretty much with an Iyonix out of the box.

"Skype's a good example of something that turned out to be a bit pointless for me. I can call Germany or Hungary for 1p/min from my normal phone by dialling a prefix. Why would I want to pay a similar amount and have to turn on a computer to do the same thing? Local calls are free anyway and I can wander around the house with my existing cordless handset."

Very nice, I'm glad for you. Other people do other things. My girlfriends' parents do the same you do, they usually phone. However, they also like a videochat. The thing is, people around here typically buy a computer to do stuff considered pretty advanced in the ROS world. One just can't pin people down to certain particular uses of a computer, however common they may be. It varies. I'll not pretend to know what everybody generally does, but I know that where I live standards are required above the level RO / Iyonix can currently provide. That's not to be harsh or anything, that's just the situation here. Still, when I find someone requiring the stuff an Iyonix or A9home can manage, I'll certainly be all over him or her suggesting to get one of the two, because I still believe ROS has great features, most of all a terrific user interface.

"I don't see why that matters to the survival of RISC OS. It's actually a potentially larger and more affluent market demographic."

OK then.

"As for preemptive multitasking: it doesn't prevent lockups. Fatal lockups happen when something scribbles all over some unprotected memory in module area or when module code gets stuck in a loop or loses track of its stack due to a bug. I'm sure someone will step in with a correction to the rule, but generally I would say: if you can't alt-break out of it, you wouldn't be able to preempt out of it either. But you generally wouldn't be able to preempt existing apps anyway which is the clincher."

OK, thanks for clearing that up for me. I've had a Windows app crash on me a while back, but it didn't hold the system until I forced it to quit which is the case with ROS. Single-tasking error or something. Other processes continued and I had the option of force quitting the unresponsive app anytime. Anyway, I'll be more careful next time

hEgelia on 22/07/06 2:20PM [ Reply | Permalink | Report ]

It is my experience too, that man people are making their own video DVDs. And most of the people I know like to download TV series via BitTorrent networks (to get them in the original language) and watch them on their computers. These videos usually come in DivX or Xvid format, but sometimes also in MPEG2, WMV9, H.264 or QuickTime MPEG4. Also to most people it would be unacceptable to use a web-browser that does not support recent versions of Flash and PDF as well as current versions of W3C standards and JavaScript. Of course all of this comes as standard with all current mainstream (Windows, MacOS and Linux) and most other modern operating systems.

The only way I can see us catching up, is by creating a compatibility layer that can execute Windows binaries as plugins for our own applications. For this we would need an x86 CPU and something similar to WINE. Then we could make use of the free codecs and browser-plugins for Windows in RISC OS.

JGZimmerle on 22/07/06 5:46PM [ Reply | Permalink | Report ]

No disrespect to anyone here, but many of the comments and observations around this are focussed at too low a level.

While I agree that that these are all important aspects, we need to look at the viability of a 'comparatively' outdated OS in the modern computing world.

While the RISCOS licence is held privately, it is much more limited in its ability to develop and modify the OS for the desktop market, as I beleive has already been pointed out. Equally, whilst ever RISCOS is dependant on the ARM platform, you are limiting development even more. Is the real issue here that people like the ARM architecture as much as the OS, or are they just reluctant to seperate the two?

Open sourcing RISCOS is just the start of a process that would require a clearly defined business model, a set of key milestones and an initial set number of deliverables.

It would provide the following obvious advantages; 1. Easier for developers and OEMS to adjust the OS to suit their needs; Ceratinly RISCOS is good for embedded devices, but opening it up allows its application in other markets, certainly in developing nations where the Windows OS has not been as invasive. 2. More diverse development is much more likley to produce the functionality we all want, DVD Burning, a good browser etc. 3. The most impartant thing this would do is increase the OS's visibility in the marketplace, which if managed and promoted well, would give more input into the OS from a financial, development and community standpoint. Currently the OS is hidden away behind a limiting licence and an architecture dependancy.

Some or all of the above points link in to Peter Wild's original post, and as already mentioned it is just the first stage in a long and potentially risky shift to a new business model for two of RISCOS's primary players. If they can't even get to stage one (or develop a viable alternative model), then the OS will eventually dissapear.

I for one hope this doesn't happend and that ROL and Castle get the ball rolling soon, before its too late.

JDC on 22/07/06 7:20PM [ Reply | Permalink | Report ]

In reply to Adam:

"I just want file compatibility, that's all. There's a lot of Mac use around here, so QuickTime compatibility is a must. I find this is a nice player for several platforms, which could someday be ported to ROS perhaps - [link]"

According to Peter Naulls it would run too slowly on RISC OS.

MikeCarter on 22/07/06 7:33PM [ Reply | Permalink | Report ]

And Peter is right. Why are there so many people who just don't (want?) to see that ARM processors are holding us back.

JGZimmerle on 22/07/06 10:18PM [ Reply | Permalink | Report ]

In reply to Julian:

The new generation of ARM processors are quite capable in the area of multimedia. The newer XScale models have WMMX which can be seen in action on recent Pocket PC models. DIVX at 640x480 run smoothly on those platforms without skipping frames. Texas Instruments even added a real DSP to offload stream processing and NVidia is now stepping up to offer special video chips for ARM processor that can handle both 3D as MPEG 4.

ARM has also kept its focus on the area of multimedia and has built the CORTEX A8 which has multimedia support and runs about 10 times faster than the Risc PC I am typing this on..

It is that the hardware developers like Symtec and Castle need a lot of time to develop hardware and adapt RISC OS for that platform.

If you would look at Toradex you would find that a complete ARM platform with XScale PXA270 at 520 MHz only costs about 100 euro's. That is because Windows CE comes with a platform builder. The OS can be configured for a specific platform and that helps developers of new machines tremendously. On top of that Microsoft has a geniric platform emulator on which software and hardware can be emulated so that developers can test their configuration before it is built..

These platforms also support Linux. Not because it has all those tools that Microsoft provides for developers but because it is opensource.

One of the biggest blocking issues in the development of RISC OS seems to me that there are plent ARM based platforms with up to date hardware but noone can use it in conjunction with RISC OS.

Either the platform gets an opensource HAL which would allow other hardware to be used or the entire OS should become opensource.

Remember that the GUI and all other desktopcandy is not to OS. These can still all remain proprietary. For example if Castle would openup ist HAL entirely (provided it caters for all the hardware access in general) there would be a chance that we could port this to many other ARM based platforms.

Just off the record, I have a Netwinder on my desk for more than 7 years now. It has a StrongARM but otherwise pretty much a PC like internal structure. Linux supports most of this hardware so it runs Linux. It could have ran RISC OS if the hardware was supported but unfortunately it is not... To try and run RISC OS software I have been working on riscose under Linux but there was too much work to be done for that to work. The basic structure for running RISC OS apps natively has been added to the Linux kernel and is still there after all those years.. Unused I am sorry to say.

For both RISC OS Ltd and Castle the licensing of the RISC OS Desktop could remain and their profits could even go up...

Just a thought though.

Jan Rinze.

JanRinze on 23/07/06 11:53AM [ Reply | Permalink | Report ]

JanRinze: "One of the biggest blocking issues in the development of RISC OS seems to me that there are plent ARM based platforms with up to date hardware but noone can use it in conjunction with RISC OS."

So true! It's all very well people name-dropping new ARM variants if none of them will ever get used with RISC OS. What you'd most likely see with an open source RISC OS is increased activity around the portability of the operating system - that's a classic activity in many open source communities which would increase the potential for wider usage of the operating system, even if portability to other architectures wasn't immediately addressed. After all, the Linux kernel was initially considered to be non-portable (or not suited to portability), but now it runs on more stuff than most people care to think about. With more attention on RISC OS from outside the commercial actors of today, there'd at least be a chance of a similar phenomenon.

guestx on 23/07/06 4:02PM [ Reply | Permalink | Report ]

JDC & Julian>But guys it may be better to support one architecture really well than support several *badly*. The most successful OS vendor in the World (Microsoft) once offered NT on several platforms (MIPS/ARC; DEC Alpha as well as traditional x86), eventually they dropped the whole notion and NT's successors (2000 and XP) are x86 only.

Not many OS'es these days have as much of their Kernel and supporting layers written in Assembler as RISC OS does. The reason why RISC OS is *in any way* as responsive at embarrasingly low clock rates as it is is *precisely* for that reason. As Jan Rinse points out there *are* more capable ARM's out there than there were - and more on the way. Now is *not* the time to dissapate all the effort in supporting a diverse range of non-ARM CPU's when more suitable hardware is arriving.

AMS on 23/07/06 4:04PM [ Reply | Permalink | Report ]

@AMS

The point I was trying to get across there was that making the OS available on multiple platforms increases its visibility, your advertising to a wider audience the benefits it can bring. Its then the platforms (and their communities) which will decide if the OS is viable or not.

Indeed there are so many other OS's (not including Windows which doesn't need to, I'll come back to that later), that are out there expounding their virtues, that if RISCOS doesn't get in the mix it will be burried.

We know we have a good OS, for a variety of reasons, now we have to tell other platforms that we've got a great OS and get them to buy into it. It's likley that it won't take off on many of these platforms, but the more we can get it onto, the more the OS gains marketability, development and support.

As for MS dropping NT support for Alpha CPU's, that's the name of the game, if it doesn't sell, then you cut your losses, market shift is inevitable, and is at least in part responsible for the troubles of Acorn back in the 90's.

You have to move with the times. RISCOS needs a new gameplan.

JDC on 23/07/06 5:06PM [ Reply | Permalink | Report ]

@JanRinze:

640y480? You must be joking. We live in the HDTV age. Most of the videos I download from the net are at least 1280x720, many are 1920x1080. You are right, that ARM is developing faster CPU cores, but the resulting chips are still not very suited to desktop computers and still achieve only a fraction of the performance of current x86 and PowerPC processors. And usually it takes ages between the availability of new ARM cores and the the availability of complete off-the-shelf chips with these cores. And quite often these chips have only stripped-down versions of the fast cores, without FPU, but with many specialised extensions we don't need.

10 times faster than a StrongARM RiscPC is still only about a quarter othe speed of a current PC. And the PC hardware is available today, but the ARM hardware will take a few more years to reach our market, by wich time the PC hardware will be about eight times as fast as the Cortex A8.

@AMS: No one is suggesting to drop ARM support and move to another architecture entirely. Windows' way of offering support for multiple hardware-platforms was poorly designed. What I would like to see is a version of RISC OS that supports multiple architectures. There are several ways to do it, wich have been discussed on drobe before. For applications we would need some kind of multi-platform-binary support. The easiest way to do this is to use GCC to compile different !RunImage files for different architectures and name them !RunImageX86, !RunImagePowerPC, etc. For applications that can not be recompiled, we would need an ARM emulator in the OS. This way we could keep the fast, hand-coded ARM assembler in most places for the ARM version of the OS, while having C versions of the HAL modules wich interface to a Linux kernel for all other architectures.

JGZimmerle on 23/07/06 5:28PM [ Reply | Permalink | Report ]

(nt)

gdshaw on 23/07/06 6:23PM [ Reply | Permalink | Report ]

In reply to Julian Zimmerle:

The method you propose for supporting multiple architectures is not scalable: everyone takes a hit (in terms of disc space, download times and so on) whenever you add a new one.

The only successful models I have seen that manage a significant number of architectures concurrently are Debian (multiple binary distributions generated centrally, user downloads one) and Gentoo (single source distribution, user downloads and compiles).

A further benefit is that you would be able to support sub-architectures (ie. optimised for ARM 6, StrongARM, XScale etc.)

(Apologies for empty message above due to finger trouble.)

gdshaw on 23/07/06 6:32PM [ Reply | Permalink | Report ]

In reply to gdshaw, there already is a successful model of supporting multiple binaries for different architectures - the Mac. These days, it's commonplace for a Mac app to have both PowerPC and x86 images in the same executable file. This was also the case on NeXTStep which also could support SPARC and PA processors.

thegman on 23/07/06 7:17PM [ Reply | Permalink | Report ]

In reply to Garry Taylor:

I thought someone might mention that, which is why I said 'a significant number of architectures'.

Yes, if you only intend to support two or three then you can get away with a naive implementation, and if (as in the case of the Mac) you are only using it to migrate from one platform to another then you will never have more than two or three.

The considerations that apply if you intend to support a significant number of platforms on a long term basis are completely different, and simply putting a binary for every platform in every application would be an awful precedent to set for an operating system which prides itself on being efficient.

gdshaw on 23/07/06 8:00PM [ Reply | Permalink | Report ]

Oh come on, it's not like the size of application executables is in any way a relevant factor, considering todays HDDs and broadband-internet. And with stuff that's for download the vendores could always offer separate downloads for different architectures. Modern applications quite often have more data-files than executable code. The approaches of debian and gentoo are not open to us, unless we want to loose all commercial applications. Software vendors will not want to open up their source code.

JGZimmerle on 24/07/06 00:16AM [ Reply | Permalink | Report ]

In reply to Julian Zimmerle:

"It's not like the size of application executables is in any way a relevant factor"

That is a value judgement with which I for one disagree. What concerns me most are the multiplicative and open-ended nature of the overhead, especially having seen some of the difficulties Debian have experienced as their number of architectures has increased (when their model only impacts on the server side).

We are already afflicated with a number of design decisions which were, shall we say, unfortunate in the long term. Leaving aside the issue of whether we have a long term future, do we really want to be adding new ones?

"The approaches of debian and gentoo are not open to us, unless we want to loose all commercial applications"

So far as the basic Gentoo model is concerned (by which I mean compiling locally), I agree. Debian is rather different, as it does not require the release of source code, and I see no technical reason why it is necessarily unsuitable for proprietary software.

gdshaw on 24/07/06 01:03AM [ Reply | Permalink | Report ]

The debian model of having big distributions for different platforms is unsuitable, because it would require all software vendors to agree on a single distribution method. If they wouldn't, then we would get the same situation as with WindowsNT for alternative Platforms, namely most software being only available for one of the platforms.

And it adds even more overhead, because then all the data-files wich make up the bulk of the total size of most applications, would all be duplicated. Take OvationPro (incl. some plug-ins) for example: total size on my HDD: 4638 KB; size of all ff8 files: 848 KB.

And with commercial software, some vendors would probably start to take money for each platform-version, so users would have to buy all their existing software again, every time they get a new RISC OS computer that uses a different processor than their previous one.

JGZimmerle on 24/07/06 02:26AM [ Reply | Permalink | Report ]

In reply to Julian Zimmerle:

"The debian model of having big distributions for different platforms is unsuitable, because it would require all software vendors to agree on a single distribution method."

Not true. Debian systems do not require that all software be installed using a single method.

"And it adds even more overhead, because then all the data files ... would all be duplicated"

Only if you package them naively, and even then only on the server (and any distribution media etc.)

I'm not saying that there isn't a point to be made here, but at the moment you are arguing it from a number of false premises.

gdshaw on 24/07/06 02:53AM [ Reply | Permalink | Report ]

t has been said here we have very good user interface and I agree but these are moving on in the outside world. These are just 3 (1 I flagged before) one thing in common is that they seem to need lots of processing power, and that is something we have not got, so how do we get it? Have a look, if you can. You will need a non RISCOS machine. [link] [link] [link]

Pete on 25/07/06 2:54PM [ Reply | Permalink | Report ]

Those all look fancy but are they any more productive? They still seem to be getting the fundamentals wrong, e.g. the Sphere UI demo scatters the windows at all sorts of angles and sizes, but still leaves the menu bar at the top of each window where they are now even more awkward to get to than in a traditional 2D desktop. The Bumptop demo doesn't show how you get your pretty looking files into and out of applications - hopefully just by dragging, but I wouldn't bet on it, more likely to be some sort of tedious file open dialogue box. Though of course it will be in 3D and bounce onto the screen in an attractive way. Which is what really matters obviously.

helpful on 25/07/06 3:31PM [ Reply | Permalink | Report ]

I am sorry to tell you that 99% of the postings above are a waste of time. Peter Wild himself told a ROL shareholders meeting that Castle cannot make RISC OS open sources as their contract with Pace when they took RISC OS over specifically excludes Castle doing so. Then there are the other companies that have rights to RISC OS who's contracts grant them rights with undertakings by the grantee, this includes RISC OS Ltd, but even if somehow both Castle and ROL agreed to it, then you would have to get Pace to agree, also E14 Inc. and possibly a couple of other companies also have rights. The likelyhood of getting all the necessary people to agree is I believe nill! p.s. Someone did post a while back apparently on E14 Inc's behalf to say they have no interest any more in RISC OS, but somehow I don't think that an anonymous newsgroup posting removes the legal rights and obligation of existing contracts.

chrisevans on 25/07/06 4:57PM [ Reply | Permalink | Report ]

What have they got to lose if they just open sourced Risc OS witghout leagal consent. Probably not much. Not many (if any) companies actuall value Risc OS.

nx on 25/07/06 5:02PM [ Reply | Permalink | Report ]

The bumptop one just irritates me every time I see it. I don't know why - but you may have hit on it in your comment; other than how to organise files, it shows you very little else. Perhaps because they haven't thought the idea through to anything more useful, such as how to actually get some work done on it, rather than spending all day throwing files around the desktop.

The sphere one, just looks like a virtual desktop which wraps around at the edges, and to make the name fit, let's make it look curved and pat ourselves on the back for being clever and original.

The touch screen one, though, I noticed a couple of things which I did think were quite good: specifically, re-sizing and scrolling/rotating. To shrink an image, the screen was touched in two places on that image and the fingers squeezed together (and the reverse to expand it). A similar technique for scrolling and rotating - touch a point, drag your finger(s) around and the document is scrolled or rotated. That worked because it's like you're saying "I want this bit (points) to be here and that bit (points) over there."

The only thing was, when it was all pretty patterns near the start, I couldn't help but think of Jean Michel Jarre performing one of the Rendezvous tracks with the laser harp.

VinceH on 25/07/06 5:11PM [ Reply | Permalink | Report ]

@gdshaw: What it comes down to, is that you either have to distribute source code (not an option for most commercial software), distribute some Java-like byte code (kind of defeats the point of supporting faster CPUs) or distribute binaries compiled for the target platform. Now I made a very efficient (you can only keep the binary for one target platform, if you're so concerned about used HDD space) and easy to implement suggestion for the third option. You dismissed it, so what is your suggestion?

JGZimmerle on 25/07/06 7:42PM [ Reply | Permalink | Report ]

In reply to Julian Zimmerle:

"So what is your suggestion?"

Personally I would go for a Debian-like model, perhaps using a package manager like RiscPkg. But then, perhaps I'm a little biased on that count :-)

As I said above, this does not mean that everyone has to distribute source code: commercial companies could distribute binary-only packages if they so wished, either on disc or from a password-protected website.

(BTW this isn't an entirely academic discussion: I am seriously interested in methods that could be used to efficiently distribute software optimised for different sub-architectures such as ARM6, XScale and so on. If anyone has any other ideas then I'm listening.)

gdshaw on 25/07/06 9:47PM [ Reply | Permalink | Report ]

But that is only a way to distribute the binaries, basically an alternative to offering downloads from a webpage. It may also solve dependencies austomatically, but but it does not really offer the fuctionality of multi-platform applications for RISC OS, where one could simply download a RISC OS application or copy it from a CD and it would run on any RISC OS platform. Also many RISC OS users do not want automated application installation with dependency checking, because they want to have full control over the installation process or like the simplicity of only having to copy an application onto their HDD to install it. To make such a system mandatory would alianate many RISC OS users. I don't think we can afford loosing any more.

JGZimmerle on 25/07/06 11:42PM [ Reply | Permalink | Report ]

to Chris Evans: no doubt you are absolutely right in your remarks about the legal position of RISCOS, but you have not addressed the central question raised by Peter Wild, and also by many of the above posters, which is how to revive and re-energise RISCOS. Or would you maintain that the status quo is satisfactory?

bucksboy on 26/07/06 10:54AM [ Reply | Permalink | Report ]

bucksboy on 26/07/06 10:57AM [ Reply | Permalink | Report ]

I don't see a problem with users buying different versions for different platforms. Commercially. which is the most likley arena for this to happen in, it makes sense especially if your wanting to support not so popular/common platforms (such as ARM).

JDC on 26/07/06 1:07PM [ Reply | Permalink | Report ]

Excellent point George.

Perhaps Chris should've considered that the other main subject matter, next to Open Sourcing ROS, is how to revive it for the desktop - which is, ofcourse, where most of us care about it. Open Sourcing ROS could be a realistic way of achieving that.

It seems fairly obvious to me that, at its current path, ROS will end up as an OS for embedded purposes, with perhaps the occassional development spin-off for the desktop market. I have one simple reason for that; The desktop market is too small to sustain a software (or/and hardware) business with. Oh, another reason - just look at where A6's and Castle's primary focus is at. As for ROL, well didn't Paul M. say at Wakefield it'll be a couple of weeks before Select 4 will be ready? It's been a couple of months since Wakefield. The word of ROL has become virtually worthless and, as a company, I can only see them becoming dependent upon Advantage Six. I'd rather not think about how many Select subscribers they have left. Perhaps they can have a future if Ad6 commissions them certain software development tasks, which they may release in one form or another for RO4 users.

Without regular news and at least some visible development, this desktop market will wither away until just a handful of ROS enthusiasts are left. Is this what Castle/ROL/AD6 want, since it seems they are surely allowing it to happen faster than ever?

hEgelia on 26/07/06 1:08PM [ Reply | Permalink | Report ]

chrisevans:

"Peter Wild himself told a ROL shareholders meeting that Castle cannot make RISC OS open sources as their contract with Pace when they took RISC OS over specifically excludes Castle doing so."

The statement was true when I made it, around two years ago. The legal position has changed substantially since that time. I am fully aware of the various legal agreements and don't believe there are any insurmountable legal barriers to Castle making RISC OS open source.

"but even if somehow both Castle and ROL agreed to it"

I'm not going to open up the ROL/Castle debate here; my original post set out the options I see for ROL. I would sincerely hope ROL would get behind and contribute to any open source initiative for the good of the whole community. If ROL's management have the acumen and vision to embrace it Open Source RISC OS, presents a real opportunity rather than a threat to ROL. The same also extends to ROL’s customers, such as AD6; given their interest in the embedded space Open Source should bring them many benefits.

If ROL's only response is to call foul I'm sure they will be rightly vilified by the user base already frustrated by their failures to deliver and broken promises. Without an initiative that brings some impetus to merging the current OS branches and invigorating the whole solution it is dead anyway and, by implication, ROL along with it. I believe everyone getting behind an open source initiative is the only way to bring harmony to the RISC OS community and promulgate its use.

To clarify a couple of things from my original post, in response to comments by various posters:

1. I’m not necessarily advocating Open Source RISC OS should be free. There will be costs involved in hosting it and managing the feedback of revised sources etc. I would hope there will be some exceptionally good, RISC OS savvy people who get involved in this process to ensure the existing reputation for quality, developed within Acorn and Pace is maintained. However, any charge for its use should be minimal to encourage people to use it and feedback new works for the common good.

2. There has been several comments on moving Open Source RISC OS to another platform. I wasn’t trying to suggest this was a way forward, only highlighting the fact that its ARM only characteristic is a big barrier to some commercial projects. If other barriers can be lowered by an Open Source project, potential customers will be more amenable to accepting this limitation.

Personally, I feel the technical and commercial barriers to moving RISC OS to another processor architecture are enormous and wouldn’t even consider such a course of action.

The above views are my personal opinions, and not necessarily those of Castle or other companies.

PeteWild on 26/07/06 2:46PM [ Reply | Permalink | Report ]

We would not have to move RISC OS to a new CPU architecture, we could also produce a cheap PCIe expansion card for standard PCs with an ARM processor on it and use that for executing ARM code, while the x86 (or PowerPC, or whatever) is used for things that the ARM can not handle.

JGZimmerle on 26/07/06 4:49PM [ Reply | Permalink | Report ]

(nt)

nx on 26/07/06 5:55PM [ Reply | Permalink | Report ]

@Julian

That's potentially workable, but introduces a hardware cost. The cost would have to be right for commercial buy-in. It would also mean opening the box to install the card, which may seem like a minor thing, but is something you don't have to do for other OS's so its a complication (and requires an above average skills set) and therefore a cost (however small)

JDC on 26/07/06 9:16PM [ Reply | Permalink | Report ]

Maybe it could be done as an external firewire device. Firewire offers direct access to the host-systems memory. And Firewire800 is quite fast. It would be an effective dongle. RISC OS could be free and open-source, but to run it, one would still need the hardware.

JGZimmerle on 27/07/06 00:06AM [ Reply | Permalink | Report ]

And it could be cheap. It would need nothing more than an ARM and a Firewire controller. Maybe there are even ARM chips wich have an integrated Firefire controller. All the things that make RISC OS computers so expensive would not be necessary. Even the drivers would only have to be generic ones, interfacing with a Linux kernel on the PC.

JGZimmerle on 27/07/06 00:21AM [ Reply | Permalink | Report ]

Personally, I see the best situation being an open source RO with the GUI remaining closed source. The GUI is one of the main reasons (if not #1) people are still using RISC OS, so to me it seems silly to allow other platforms to steal our main advantage. It would leave both OS companies in a better position, because they could still be able to sell the full desktop version of the OS, while it suits ROL quite well. With their lack of resources they are probably better off concentrating on features which can be marketed easily, rather than "LD / ST T in abort trap mode" etc., which is pretty irrelevant to many, while allowing others to do core OS work. They seem to be pretty good at implementing new GUI stuff (better than hardware support?).

I guess the only drawback with this method is that potential development resources might be hindered, since people still have to pay to use the Desktop OS. But I think ROL would be happier receiving some guaranteed money; less of a risk for them maybe?

"Maybe it could be done as an external firewire device." That's a really good idea. It can't beat the free shipped CDs from the Linux world, though

timephoenix on 27/07/06 06:10AM [ Reply | Permalink | Report ]

timephoenix: use of the word "steal" is entirely unappropriate.

Surely if the RISC OS GUI is the best thing, getting it ported and available on other platforms would be a good thing.

If you're seriously concerned about Microsoft running off with it, there are numerous open source licences (e.g. the GPL) which would prevent that.

Open sourcing RISC OS isn't a magic bullet, unless it's an attractive and appealing proposition for OS tinkerers to get involved; without a GUI I imagine that set of people would be reduced further.

Jaffa on 27/07/06 09:08AM [ Reply | Permalink | Report ]

Personally I don’t see RISC OS going much further in the desktop market (as in the amount of users). I believe its future lies in embedded systems such as; set top boxes, interactive displays, home media systems, car computer systems, etc. But other than that it’s a waste of time and money focusing on the any other market.

MikeCarter on 27/07/06 09:16AM [ Reply | Permalink | Report ]

JGZimmerle: thats a bad idea thats been put round since the days of the 486 PC and is a worse now than ever - you'd be laughed out of town for having a dual core PC sitting there, but actually running your software on the same chip as used in the ADSL modem. Using the type JIT technology that Apple uses rather than the simple version in Virtual Risc PC, you can emulate in software on an x86 far faster than any current ARM chip. If you want some sort of dongle to stop piracy then put the code on a USB pen and lock it to the serial number on that rather than the specific PC so its portable (an idea rejected by VA however).

But most of this discussion is making RO open source, so then you don't need to lock it to anything, you want as many copies to be made as possible. But emulation is the ultimate dead end, to have any future you have to follow what Apple has done (twice) and move to running native on the new architecture. Even if the OS was open sourced, it would need more than just porting to run on something as diverse and rapdily evolving as x86 hardware, i.e. it would need to sit on top of an existing kernel and driver model as Mac OS X does on BSD, which would be a huge task. An even bigger task would be moving applications to a new archictecture as without this, they would still all be emulated giving no real advantage. Even if its done the huge development work is just to get a working native system, and not producing anything new.

You have to think beyound fancy ideas, such as if the OS were just open sourced and merged everything would be ok, or if we ran on cheap fast PC hardware everyone would be happy - they wont be. RISC OS users dont care about the type of OS licence, they don't need to know if its CMT or PMT as long as it works, they are even willing to buy expensive and relatively slow hardware, but what they do care about is having RISC OS applications that allow them to do the same thing with a computer as they see everyone else doing. If the development community goes off and spends a few years tinkingering with an open source OS, and not writing any applications, then there will be no users left to care.

It comes down to this, either we make the OS and hardware that we've got do something useful for people by writing applications, or if you want a substantially different OS running on other hardware then a clean break and moving to Linux + ROX or Mac OS X is going to waste far less of your life than trying to drag along any of RISC OS source.

druck on 27/07/06 09:52AM [ Reply | Permalink | Report ]

Well, I agree that the preferred solution would be a complete open-source port to x86, using a Linux kernel. Only if this turns out to be impossible or unpractical, I would suggest the ARM-in-a-Firewire-box solution.

I don't agree about your suggestion that either of the two would prevent new applications from appearing for RISC OS. I think once we had the more powerful hardware available, we could finally run ported applications like Firefox at decent speeds.

JGZimmerle on 27/07/06 10:31AM [ Reply | Permalink | Report ]

Julian>I believe it was Peter Bondar who originally suggested putting a StrongARM on a PCI card and fitting it to a PowerPC (which I suppose if Acorn were sidling up to Apple at the time might make sense to them). It was, however, resoundly rejected by the user base. I'd also remind you that apple have gotten out of the PowerPC business so if we *had* followed that path we'd have been right properly stuck (also the RPC, Iyonix, Omega and A9 would never have happened....).

Now you're suggesting the same approach....

The *best* option is to keep RISC OS running on ARM (I'd point out that Arm have now announced support for 65nm and 18nm processes - this would represent a substantial speed up). Now is *not* the time to jump ship.

As to many of the suggestions made here to "open source" RISC OS inevitably "Linux" is presented as part of the solution. Now I don't have an issue with Linux (I have my own RH Linux machine at home) - but I don't want to see Linux get more users at the expense of RISC OS - which is what a lot of the suggestions here seem to amount to.

Besides (as David Ruck pointed out) "they [RISC OS users] do care about is having RISC OS applications that allow them to do the same thing with a computer as they see everyone else doing". Point is that Intel and Microsoft and the MPAA/RIAA seem to be pushing DRM that (effectively) would prevent anything *other than* Windows from using HD-DVD or other (modern) multimedia. That means whether you'd use Linux or RISC OS you're in the same boat - eventually you just simply *won't* be able to view Videos or play music on anything other than a Microsoft OS box.

So does an "open RISC OS" need Linux or would it work just as happily on top of Windows ? If the options of doing modern Multimedia requires WINDOWS what do you think the likely outcome is going to be....

A closed source RISC OS might be able to *license* the DRM technology (a more likely outcome than Linux getting it I would have thought) - and with faster ARM hardware on the way it may also become a technically possible outcome (remember MS are trying to get into the Consumer Electronics (CE) market - it might be possible to classify RISC OS machines in those terms especially in the embedded CPU space). It *might* be possible to use MS DRM *within* a RISC OS environment running on faster native ARM hardware.

Yes I do know DRM is evil - but when people click on a video and it won't play they will simply buy a platform that *can* run it - and before long that may just be Windows. Pushing for ROS running on Linux might simply result in RISC OS running on Windows - an option I for one would certainly not be keen on.

AMS on 27/07/06 1:46PM [ Reply | Permalink | Report ]

@AMS: I am NOT suggesting to abandon ARM. How many times do I have to say that? I am saying, that we should find a way to utilise other CPU architectures as well. That is why I suggested to have multiple !RunImage files, each one compiled for a different architecture. How these !RunImage files would be executed, would be up to RISC OS, wich in turn would use different methods, depending on the hardware configuration. There could be a way to execute !RunImageX86 files on RiscPC's second processors (useful for FP-intense software?) and another to execute the same !RunImageX86 file on a PC with an ARM attached via Firewire or PCIe and another to execute it on an Iyonix with an internal x86-PCI-card. The same principles could be used with PowerPC and other CPUs. It could also be used to speed up RISC OS applications in emulated environments.

JGZimmerle on 27/07/06 5:55PM [ Reply | Permalink | Report ]

@AMS: And yes, we should use the Linux *kernel* to get rid of all the very expensive and time-consuming driver development. This way we would *not* loose RISC OS users to Linux' desktop environments. It would still be RISC OS, but the hardware-abstraction would be done through the Linux kernel. There will also be Linux software to play DRM protected files, or to extract the content out of them. There are special provisions in the IP-laws of many countries, that allow the cracking of DRM systems on platforms that have no other way to play these files.

JGZimmerle on 27/07/06 6:08PM [ Reply | Permalink | Report ]

In reply to Julian:

RISCOS could be run under Linux. There has been a group of people involved in that. I was one of them. There even is a special flag for kernel compilation called RISC OS personality. It caters for using SWI calls in 26 bit mode on a StrongARM. Unfortunately the route involved writing opensource modules to implement SWIs (see riscose). This proved too much a task. Also when Chris Rutter died the project had lost its most motivated member. There is no reason why 32 bit modules can't be used on ARM Linux. Nor is there a restriction to running RISC OS 32bit or 26/32 bit neutral applications except for the lack of an AIF loader. On top of that we would have to fill in the gaps where hardware is addressed. A port of the Castle HAL could simplify that task and maybe even allow RISC OS to be run natively in a VMWare like environment.

For the time being I believe that there are 2 things we need to keep an eye on. Firstly how can we utilize new hardware if the companies that sell RISC OS don't support them? And secondly how can we 'upgrade' RISC OS to more contempory standards (including user applications).

Using Linux as a bridge might help to get both access to 'compatible' new hardware as well as to have an environment in which there are many solutions to currently missing functionality.

I am not saying we should all rush-off and try to build such a beast but it could be a direction in which RISC OS could keep its specific character and all of its commercial value next to the benefit from progress made by a huge group of programmers.

Just my thoughts though..

Jan Rinze.

JanRinze on 27/07/06 7:40PM [ Reply | Permalink | Report ]

"...use of the word "steal" is entirely unappropriate."

Fair enough, perhaps that could have been worded better. But what I don't want to happen is for an improved ROX desktop to be released and native RISC OS to die. The good RO software, along with the GUI, is all that keeps me from migrating to Mac.

timephoenix on 27/07/06 10:11PM [ Reply | Permalink | Report ]

@JanRinze: To me your suggestion seems to be the best way forward. The big question is: How do we get Castle and RISC OS Ltd. to sign up to such a plan?

@timephoenix: I don't think anyone seriously interested in RISC OS wants that.

JGZimmerle on 27/07/06 11:16PM [ Reply | Permalink | Report ]

I don't see how it is realistic to merge two OS source trees that have diverged over so many years. It was a very difficult and long drawn out process trying to merge some of the mess caused by the ANC and ART branches of RISC OS diverging over a much shorter timeframe.

I agree that open source is by no means a panacea - there are still many ways of killing it all off: e.g. lack of quality control, choosing an inappropriate licence such as GPL (unless it's dualled with modified BSD or similar), the remaining developer(s? ;-)

getting bored and just going away.

I would say that the idea that other developers of other OSes would pinch ideas from the source is loopy, to be honest. RISC OS *in itself* has nothing to offer that other OSes don't already have in a far more sophisticated form already (for many years). The quoted example, the Window Manager, is a huge amount of ARM assembler that achieves relatively little in terms of functionality. It's a half-way house between a rectangle manager and a GUI environment and it does neither particularly well. It's major plus point for users is that it enforces various common traits in applications (icon bar positioning, window furniture behaviour), whilst that is also a weakness (lack of flexibility (cannot put the icon bar down the side of the screen instead of the bottom, for example), cannot customise behaviour of furniture easily, cannot use the furniture within your application). Thus you can get inter-operable applications with a common look-and-feel and that helps users, but ...

Users don't care about the inner workings of the OS directly - they *do* care about it indirectly, even if they don't know that they do. Users cannot and should not be expected to understand the inner workings of the OS. Consequently, I don't believe they realise just how awkward it is to write Wimp applications (particularly those that interact with the hardware) given the architecture of the OS.

stewart on 28/07/06 00:49AM [ Reply | Permalink | Report ]

Stewart:

Why is GPL an inappropriate license? It means any developments stay GPL, out in the open on a level playing field. BSD would mean the code could disappear into other commercial products.

markee174 on 28/07/06 07:50AM [ Reply | Permalink | Report ]

markee174: I'd say that's a perfectly good reason for saying that GPL is an inapproriate language. It'll put some people off.

Personally I think things should either be completely open and free, or closed. The GPL is neither.

If there was some hypothetical open source RISC OS, who's going to develop it? RISC OS has a developer shortage as it is, and I can't imagine that many people jumping in to help out. Maybe a few ex RISC OS users who'd be curious to have a look in it, and if we're lucky one or two might stay.

The whole question about opening up some or all of it (whether or not it would be possible) seems to me to miss the point. Stuart Brodie said "Users don't care about the inner workings of the OS directly", they just care about what they can do with it, and whilst it's on the hardware it's currently on it won't ever be able to compete, even if every single other bit of wishful thinking about RO was achieved.

It needs to find a niche, not be open sourced.

SimonC on 28/07/06 09:53AM [ Reply | Permalink | Report ]

Julian>Yes I have noted that you have *not* suggested that we don't use the ARM. Your suggestions would, however, mean that the concept of a desktop ARM based computer running RISC OS natively would cease (IMHO), which I think would be a bad thing.

SimonC>"whilst it's on the hardware it's currently on it won't ever be able to compete"

Yes - but the solution is not to say let's decamp to an x86 based Linux box that either emulates (or better still) runs an x86 binary (translation) of the ARM RISC OS code. Thing is that there are *faster* ARM's on the way, with large caches, higher clockrates and smaller geometries (all which boost performance). ARM *has* to up their game - now, IMHO, is *not* the time to leave.

AMS on 28/07/06 6:58PM [ Reply | Permalink | Report ]

SimonC: "It needs to find a niche, not be open sourced."

Unfortunately, that niche is unlikely to be any form of viable, long-term desktop market, AFAICS - and with commercial companies (quite rightly) focused on the embedded bits (or whatever) which make them the most money; any desktop development is most likely going to be an afterthought.

If *nothing* else, an open source RISC OS would allow the wider adoption of Select's new APIs and thus encouraging their use by more developers: making it easier to write new, funky, applications which attract the users.

But without a *radical* sea-change in terms of architecture (both software and hardware) and support, I can't personally see any viable desktop future for RISC OS. At least open sourcing it would mean it wouldn't have to be a successful *commercial* OS to continue, just be interesting to at least one developer and/or one user.

Jaffa on 28/07/06 8:53PM [ Reply | Permalink | Report ]

AMS: "Thing is that there are faster ARM's on the way [...]"

But who's going to design and build the hardware to house such an ARM for a desktop computer likely to sell in the hundreds? No-one can *seriously* argue, surely, that hardware is what sets a RISC OS computer apart? Yes, they can be quiet and small because of the hardware; but so's a Mac Mini and that's got a dual core x86 chip in it.

If it's the software - whether applications or GUI - let's find a way of getting that to run on commodity hardware, so developers/users/interested parties have a) a lower barrier to entry to the platform; and b) can invest in furthering the software, not designing new low-end (in the real world) computers to run them on.

The fact that commodity hardware, being so much faster in terms of raw grunt, will also open up the possibility of doing all those things which no RISC OS computer can feasibly do - DVD ripping/playback; video editing; massive compiles etc. - is another reason it should appeal.

Jaffa on 28/07/06 8:59PM [ Reply | Permalink | Report ]

We can argue until the cows come home about ARM based machine vs commodity hardware and most of the comments above have merit however the licensing of the OS is, I believe, still the major obstacle.

The current situation has left us all victims of chance and the whim of a few buisnessmen, some of whom knew very little about RISC OS itself. I would relect that since Acorn's demise I've spent more on things that looked hopeful for RISC OS but which turned out to be nothing of the sort than on products that still exist. How many other people bought Psion netbooks because of the rumours of RON? How many other people bought a copy of that graphics package that simulates the effects of the inks used in particular printers (the name escapes me now but it seemed to have great potential at the time only to be unusable now due to activation requirements not to mention the bugs! - I can well understand the annonyance many people are showing towards VA's similar choice). Even issues of the number of developers are related to licensing. Why write/update an application for a platform which could so easily disappear? Contrast this with the fact that we can be very comfortable that one thing related to RISC OS, support for the ADFS format, will survive for the forseable future - it's part of the kernel.org linux distribution.

I don't mean to be pessimistic - comments over the last few weeks have given me the first glimmer of hope that more future-proof licensing is possible.

k626 on 29/07/06 08:57AM [ Reply | Permalink | Report ]

In reply to Simon Challands: The GPL may put some people off, but there are also people who won't even consider contributing to an open source RISC OS unless it is guaranteed that users will always be able to have access to improvements.

"Personally I think things should either be completely open and free, or closed. The GPL is neither."

The GPL does its job very well. It keeps the source open for the end-users, even if they don't care about that aspect of the product.

"It needs to find a niche, not be open sourced."

Hasn't anyone learned anything about the desktop marketplace since Acorn's demise? There's been this tendency to go prospecting for new niches in the hope that a lucky strike will bankroll desktop development, but this constant search for new markets seems to end with the current OS owner(s) focusing on pleasing their new/prospective clients rather than their former core market.

The rampant denial about the advantages of open source development in the RISC OS world, particularly in the case of certain companies with the responsibility to develop the platform, can only lead to a situation in which there are ever fewer locked in and embittered users, short changed by vendors who can't deliver without substantial funding from an external party. But who would want to invest in RISC OS when you have a situation like that?

Let's see what RISC OS Open have to offer.

davidb on 29/07/06 5:44PM [ Reply | Permalink | Report ]

gdshaw: "Debian is rather different, as it does not require the release of source code, and I see no technical reason why it is necessarily unsuitable for proprietary software."

You mean the Debian packaging system here, I take it, given that the Debian organisation is all about making source and binaries available together in a convenient fashion in order to honour various Free Software licence requirements. It's certainly true that various closed source vendors use .deb files to make their wares available, though.

timephoenix: "Personally, I see the best situation being an open source RO with the GUI remaining closed source. The GUI is one of the main reasons (if not #1) people are still using RISC OS, so to me it seems silly to allow other platforms to steal our main advantage."

An Apple-inspired approach? But why would people be so interested in hacking on a severely limited portion of the operating system code when the desktop is what it's all about? The main advantage is not exactly the desktop's set of features (many of which are available elsewhere in more competitive forms, anyway) but the efficiency these features have on low-end hardware. It's hard for people to "steal" the frugality of the RISC OS desktop. and they'd probably not be so interested in doing so, either, given their desire to work in high-level languages and their lack of interest in adapting an existing desktop environment for systems without a hard disk or with less than, say, 128MB RAM. But people may be interested in porting RISC OS to other architectures in order to apply that frugality to other kinds of systems - it'd still be RISC OS, albeit on Intel or PowerPC, and with the copyright heritage remaining intact. It would be a meaningless exercise to try and encourage people to port RISC OS without the desktop, mostly because you wouldn't get most of the benefits of the system in the resulting code (who wants to run something like MOS in console mode on amd64?), and partly because people can probably find more exciting embedded systems projects to work on instead.

guestx on 29/07/06 6:57PM [ Reply | Permalink | Report ]

Andrew Flegg wrote >"who's going to design and build the hardware to house such an ARM for a desktop computer likely to sell in the hundreds"

No one.

But here's the thing, the ARM based hardware can be "dual" tasked - to act as both a desktop product and as another dedicated product (e.g., NAS controller, Set top box or whatever). To extent that's already happened (with the Iyonix and A9Home). It's a matter of *keeping it happening*.

ARM hardware is *thankfully* improving - can have VFP (Vector Floating Point), can have larger secondary cache, can even have cache coherency (bus snooping). And finally even ARM is discussing migration to smaller geometries (65nm and below). All of this means considerable performance improvements *even* if the clock rates remained the same (and yes they're to increase too).

Will anyone build a box with hoards of "interlopers" whose interest seems more to promote "Linux" and "The GPL" than to support RISC OS ? I'll quickly qualify that by saying I have used Linux (even have it on one of my boxes at home) - and nothing against it - but simply suggesting that RISC OS can be "saved" by getting it to run on Linux is no more a realistic prospect than putting it on Windows.

David Boddie wrote >"The rampant denial about the advantages of open source development in the RISC OS world."

Where ? OSS can be used for apps (like Firefox) and other applications at the moment. So how come there's no open Database products, open DVD ripper/decoders. If the number of OSS apps for RISC OS can be counted on the fingers of one hand - then why oh why should Opening RISC OS change things ? The reality is (IMHO) that the problem is a shortage of developers. Yes opening RISC OS may encourage some "non" RISC OS developers to port some RO features into Linux (say) - but I can't say that that would translate as a sudden upsurge in improvements of the native platform or RISC OS itself - but I'd be happy to be proven wrong.....

AMS on 29/07/06 7:08PM [ Reply | Permalink | Report ]

markee174: "Why is GPL an inappropriate license? It means any developments stay GPL, out in the open on a level playing field. BSD would mean the code could disappear into other commercial products."

People shouting the loudest about the "freeness" of BSD are typically those who would like people to give them stuff but not to give anything away themselves; you can call it the freeloader aspect of free if you want.

SimonC: "I'd say that's a perfectly good reason for saying that GPL is an inapproriate language. It'll put some people off."

Yes, all that ideological rhetoric has really kept the big names like IBM away from Linux.

SimonC: "Personally I think things should either be completely open and free, or closed. The GPL is neither."

The GPL keeps software completely open from the author right the way through to the end-user. Any pretense that permissive licences encourage openness surely collapses under scrutiny as the first freeloading commercial entity sucks the source into their internal version control system, and spits out binaries to their end-users thereafter.

guestx on 29/07/06 7:11PM [ Reply | Permalink | Report ]

AMS: "I believe it was Peter Bondar who originally suggested putting a StrongARM on a PCI card"

Have a look on Simtec's product pages for exactly that kind of product. In fact, look here:

[link]

How about a bunch of StrongARMs on a PCI card? Don't confuse Bondar's incoherent missives for the remarks of a visionary, though.

guestx on 29/07/06 7:15PM [ Reply | Permalink | Report ]

I wrote about, "The rampant denial about the advantages of open source development in the RISC OS world."

Annraoi McShane wrote: "Where ?"

Here in this thread, and often on the comp.sys.acorn newsgroups, where educating people about licensing seems to be an ongoing task - maybe the situation is improving. It seems to me that lots of people still think that RISC OS could still have its day as a closed source consumer operating system if only someone would come along and throw money at it.

"OSS can be used for apps (like Firefox) and other applications at the moment. So how come there's no open Database products, open DVD ripper/decoders."

On RISC OS or on other platforms?

"If the number of OSS apps for RISC OS can be counted on the fingers of one hand - then why oh why should Opening RISC OS change things ?"

That's a remarkably blinkered view of things, even if we're talking about end-user applications. Loads of open source applications, with many different licenses, have been released for RISC OS over its lifetime, and a lot of this happened without the help of Acorn or any of the other maintainers of the OS. At the very least, opening up RISC OS would indicate a change in attitude towards open source. Beyond that, it would allow interested parties to improve parts of the OS that don't make money for the current vendors.

With the sources available for non-core parts of the OS, you could do even more. Think about the possibilities for improving bundled applications like Draw. You can't say you wouldn't be interested in that!

"The reality is (IMHO) that the problem is a shortage of developers. Yes opening RISC OS may encourage some "non" RISC OS developers to port some RO features into Linux (say) - but I can't say that that would translate as a sudden upsurge in improvements of the native platform or RISC OS itself - but I'd be happy to be proven wrong....."

I don't see how it can be any worse than keeping it all closed. At the moment, if you want to get developers to work on RISC OS, it seems to me that you have to find people who have had that job before then get them hired by Castle or whoever. With an open source RISC OS, anyone could work on (or at least look at) the code, including all those former Acorn engineers.

davidb on 29/07/06 8:07PM [ Reply | Permalink | Report ]

Just a reminder. Due to the way that RISC OS is organised it is entirely feasible to replace it with an opensource version whilst having an originbal (booted) version as a commercial product. This does mean that a specific module will have to be improved by re-writing from scratch. However, that might not be a bad thing for some modules.

I have suggested before that somebody (not me since I have no idea where to start) should set-up and opensource RISC OS development area. This should include applications and modules. If it fails then we are in no better position than at present. If it succeeds then the commercial developers may be then be persuaded to add their products as opensource, including the intransigent Castle and ROLtd. I for one would love to have a single location for all my non commercial RISC OS needs.

Or would could always talk and argue ourselves into extinction!

mripley on 31/07/06 08:21AM [ Reply | Permalink | Report ]

Hi Malcolm.

Is this legal without the authorisation of the company? I could use a computer analogy. EG making a macOSX clone, however, say you were to make car components that looked like another car and the workings were to mimic say a BMW 3 series. Would this not infringe the BMW rights.

Is software such a grey area. EG you can make a wordprocessor freely (eg openoffice) that reads file formats but you cant say we are going to create our own free version of Word by replacing the MS modules. Surely this is not legal. So how would it be possible to make something that looks and works like an OS in this way?

Now I do support an Open source RiscOS for this one reason........ It is not economic for me to buy an iyonix. I wont use a slow RiscPC (image analysis work) and I won't buy Virtual RiscPC since I wold need a copy on 3 machines (even though I only run one at a time).

If, however, RiscOS was open source, I could BUY, yes buy, virtualRPC and put it on 3 machines with opensource RiscOS, since I beleive the limitation is with the ROL license. This would mean I may buy other apps to go along with it. If this doesn't happen then RiscOS is dead, if not already, as an OS to attract new users, since it has a small software base.

I think we are talking with our last breath.

Cheers Bob

nijinsky on 31/07/06 12:43AM [ Reply | Permalink | Report ]

GuestX>Thanks for that URL. By the way I wasn't actually making any comment about Peter Bondar's "concept" rather pointing out that the userbase didn't like the notion of having RISC OS reduced to running on a card incide an Apple PowerPC machine.

David Boddie>Indeed I wasn't saying there weren't Database or other OSS apps available on *other* platforms (such as Linux, BSD etc.,) but rather that simply having OSS didn't necessarily mean that they'd be ported to RISC OS. Biggies (like Firefox) brought to RISC OS are rarities (unfortunately) - and I don't believe open sourcing the OS will change that - why should it ? Other OSS software (or software released with Source on ROS) does occur but usually it is the product of small groups or sole developers who *would* have produced it on RISC OS irrespective of RISC OS's open or closed status.

AMS on 31/07/06 1:26PM [ Reply | Permalink | Report ]

A more general question (maybe someone can explain this to me....)

Peter Wild has 30% of Castle's voting stock... He has stated he's in favour of open sourcing RISC OS.

Ok.

Now if this is such a good idea, and if that 30% of voting stock has such influence why hasn't he [Peter Wild] recommended it to (and persuaded) Castle to do it *already*. If he couldn't or can't - why should these shares if bought by (say) RISC OS Open or any other user group succeed where he failed ?

Just a thought.....

AMS on 31/07/06 1:30PM [ Reply | Permalink | Report ]

Annraoi McShane wrote: "Indeed I wasn't saying there weren't Database or other OSS apps available on other platforms (such as Linux, BSD etc.,) but rather that simply having OSS didn't necessarily mean that they'd be ported to RISC OS."

I thought you meant that. :-)

"Biggies (like Firefox) brought to RISC OS are rarities (unfortunately) - and I don't believe open sourcing the OS will change that - why should it ?"

It depends on whether developers feel that porting open source software to a proprietary platform is worth their time or not. People port software to Windows, but probably only because it allows them to reach a large number of new users. You can't say that about RISC OS.

"Other OSS software (or software released with Source on ROS) does occur but usually it is the product of small groups or sole developers who would have produced it on RISC OS irrespective of RISC OS's open or closed status."

That certainly used to be the case. Acorn didn't exactly encourage developers to port open source software to RISC OS (unless you had connections), so it happened more in spite of Acorn rather than anything else.

These days, the RISC OS platform isn't such an attractive target for ports or for new natively written software, so people want something in return for their efforts, and opening the source code is at least helpful in making the development process more accountable and transparent, something that helps foster trust between the OS vendor and developers.

In the past, it was recognized that you didn't have a viable platform without applications. These days, I think that the OS maintainer(s) feel that they have more to gain (or maybe less to lose) by keeping it all closed than by opening it up. In the end, we may have to wait until they have nothing to lose and, by then, it may be too late.

davidb on 31/07/06 1:56PM [ Reply | Permalink | Report ]

AMS: "Thanks for that URL. By the way I wasn't actually making any comment about Peter Bondar's 'concept' rather pointing out that the userbase didn't like the notion of having RISC OS reduced to running on a card incide an Apple PowerPC machine."

That was not the concept, use of a PowerPC chip never came in to it. Apple were talking about developing a Common Hardware Reference Platform, standardising on components such as memory, disc and I/O controllers to gain some of the economies scale seen in the x86 world. An ARM chip/daughter card would have plugged in to the main processor slot of such a machine, as an alternative to the PowerPC or anything else (MIPS, Alpha), not in addition to it.

The reason why this was unpopular amoungst the user base is that it came just after the launch of the Risc PC, which was meant to be future proofed due to the replaceable processor card feature, and moving to a CHRP based platform at some later date was seen to completely negate this.

druck on 31/07/06 3:13PM [ Reply | Permalink | Report ]

Malcolm Ripley wrote: "I have suggested before that somebody (not me since I have no idea where to start) should set-up and opensource RISC OS development area. This should include applications and modules. If it fails then we are in no better position than at present. If it succeeds then the commercial developers may be then be persuaded to add their products as opensource, including the intransigent Castle and ROLtd. I for one would love to have a single location for all my non commercial RISC OS needs."

Do you want to reimplement RISC OS from the ground up on RISC OS? Well, you could start with the Shared C Library:

[link]

"Or would could always talk and argue ourselves into extinction!"

There's always that. :-)

Personally, I think it would be better to more productive to try and influence desktop environments on Linux to take on board some of the more user-friendly aspects of the RISC OS desktop rather than build something from scratch with the same architecture.

The idea of having RISC OS applications running on another system is tempting but, if you end up having to implement all the low-level nastiness in tedious detail and all the applications are still not taking advantage of higher level libraries and tools, you're just prolonging the agony. Many applications need to be migrated to better development platforms and, if those platforms have desktop environments that are familiar to RISC OS users, then so much the better.

You could argue that opening the source to RISC OS could ease this migration in many ways, and that this would seem to work against the current OS maintainer(s). However, it might also lead to robust run-time environments being written to enable RISC OS applications to be run on other platforms. In this scenario, the OS vendor becomes a library or framework vendor instead. Surely there's a business case for that.

davidb on 31/07/06 3:51PM [ Reply | Permalink | Report ]

In reply to GuestX:

"The GPL keeps software completely open from the author right the way through to the end-user. "

That in itself is not the problem. The problem is the infectious nature of the licence, which is why many companies will not touch any GPL software at all in the first place (including my current employer and past employers). GPL software can never be part of a larger solution that is not entirely GPL licenced (basically speaking anyway, and the LGPL isn't as "lesser" as some people would have you believe either). You may believe that this is a good thing. I think it's crazy.

--Stewart

stewart on 01/08/06 10:21PM [ Reply | Permalink | Report ]

That is complete nonsense. There are many examples of larger solutions, wich use GPL software. Many GPL'd server programs are sold in conjunction with closed-source web-configuration-frontends for example.

JGZimmerle on 02/08/06 00:48AM [ Reply | Permalink | Report ]

No, that's not what I meant - in that case, they are totally separate communicating programs - which is OK provided that no parts of the server code are linked into such a frontend.

I'm sure some of the biggest open source nutters would like to extend the GPL's invasiveness to include such systems, though! :-)

stewart on 02/08/06 01:08AM [ Reply | Permalink | Report ]

Stewart Brodie wrote: "That in itself is not the problem. The problem is the infectious nature of the licence, which is why many companies will not touch any GPL software at all in the first place (including my current employer and past employers)."

The problem is really that the term "infectious" implies that, when companies fall foul of the license, it's at best accidental and at worst the fault of the license. People make conscious decisions about the software they use to get their jobs done, and many people claim that the license is "infectious" to explain away ignorance about licensing issues or negligence when people are put under pressure to meet deadlines.

Show us a license that achieves the same aims as the GPL, but isn't "infectious".

Back on topic: I can understand why you in your professional capacity might not want to see RISC OS relicensed under the GPL. After all, I suppose there are various companies that probably want to use RISC OS, perhaps in consulting projects, and who feel that it would be to their disadvantage to let their clients see the source. I would have thought that the selling point of RISC OS for some of those companies is the experience and expertise of their employees rather than the ownership of the operating system itself.

davidb on 02/08/06 12:47AM [ Reply | Permalink | Report ]

stewart: "GPL software can never be part of a larger solution that is not entirely GPL licenced (basically speaking anyway, and the LGPL isn't as "lesser" as some people would have you believe either). You may believe that this is a good thing. I think it's crazy."

Aside from various differing interpretations of things like linking and derived works, the LGPL merely requires you to propagate the rights to the covered source (that you as a developer enjoyed) onto the end-users, whether you think they'll be interested or not. If you don't think they'll be interested, and there isn't enough space on that 800K E-format ADFS floppy, then a promise to supply the source is all that's required - it doesn't even have to be your own application's source if you've been paying attention. Some people think that this is all so much overhead, which makes you wonder how they actually run any kind of business whatsoever given the relative simplicity of such matters compared to other logistical challenges in the realm of commerce.

With regard to whether you believe the matter to be crazy or not, I don't see why having access to the source code (or the right to such access without bizarre or unethical contracts between the parties concerned) is a bad thing. The majority of users on the RISC OS platform (past *and* present, just to bump the numbers up to make the point) have most likely been bitten by a lack of access to data stored in dodgy, binary, proprietary formats and controlled by discontinued proprietary software. If preventing the continual disempowerment of end-users is crazy then unlock the door to the padded cell and show me inside, where I'll surely be able to make conversation with more forward-looking executives than those the RISC OS scene has managed to produce.

stewart: "I'm sure some of the biggest open source nutters would like to extend the GPL's invasiveness to include such systems, though!"

The issue of retaining control (or more accurately, the means of control) of your information, where you've possibly delegated that control to some opaque Web 2.0 business, may not be directly addressed by GPLv3, even though many have argued for it based on various existing open source licences that do address this issue, but that doesn't mean that authors shouldn't be able to indicate in a standard way that their work should not be used to operate businesses which see such transparency as a matter of occasional discretion.

Anyway, if you don't like the Free Software family of licences, just don't use them. But with that, how about dropping the contentious terms and derogatory remarks?

guestx on 02/08/06 7:38PM [ Reply | Permalink | Report ]

In reply to Stewart Brodie:

"The problem is the infectious nature of the licence"

If you think the GPL is infectious, consider what would happen if you reused code taken from Microsoft Windows, RISC OS, or just about any other piece of proprietary software.

gdshaw on 02/08/06 9:47PM [ Reply | Permalink | Report ]

I've been reading this and other threads about the future of RISC OS and I have to say that there are a couple of things that I've stumbled upon that no one seems to mention anywhere.

First off, some people see the future of RISC OS as an embedded OS. Why would anyone here want that? I thought that the reason people stick with RISC OS is for its UI (at least that's what everyone mentions when asked what they like the most). Becoming an embedded OS will most likely mean most of the UI is discarded for a much simpler one or possibly not even an UI at all depending on the equipment it's being used in. I personally don't give a damn if the tv decoder runs RISC OS, Linux, BSD or even AmigaOS as long as it does the job. I'm never going to see anything other than a simple screen that would look exactly the same no matter what the underlying OS is. RO may be usable in PDAs or mobile phones or other such areas but seems to have a lot of caching up to do only to be on a level playing field to those that dominate this area.

Secondly (and more on topic). Making RISC OS open source isn't going to solve anything automatically as many have pointed out here. What no one seems to have realised (at least not mentioned) is the added fragmentation this could and most likely would introduce.

There is a group of people that sees RO future only on ARM hardware and a part of that group also believes that moving from assembler to a higher level language (like C or C++) would be almost sacrilege. Then there's another group that sees RISC OS as nothing but an operating system that should move to another hardware platform like x86 or whatever may be available. This would require rewriting the ARM assembler parts in C/C++ or whatever language people might prefer.

There are even two more groups that split between all the above groups, those that feel that the OS should continue to be co-operative multitasking and those that want RISC OS to become pre-emptive multitasking with multi-threading support.

Only this gives a possibility of SIX versions of open sourced RISC OS. Then there is the possibility of moving to another existing kernel such as linux, bsd or whatever else. There's absolutely no question that at least some of this fragmentation would happen so instead of two closed source versions of RO there could be 7-8 different versions of an open source OS in a worst case scenario.

It may seem that I'm totally against RISC OS becoming open source and only see the bad points in this which is not true. I would have nothing against RISC OS being made open source, I just think people need to think the whole thing through before jumping forward and assuming that open source is some automatic saviour for a dying desktop or that open sourcing is all evil and nothing good can come out of it.

Gulli on 02/08/06 11:45PM [ Reply | Permalink | Report ]

To answer some of Gunnlaugur's concerns. Any project has to have the scope pre-defined otherwise it just grows arms and legs and goes off in all sorts of directions. Such a scope would define the boundaries within which any RISC OS development occurs e.g It will remain a co-operative multitasking system. Now I know agreeing on such a scope with an Opensource project could be difficult but there is a way forward: we ask the users. However, before we even contemplate changing the scope of the current RISC OS we should bring it into the 21st century. This means the OS stays as is but has all shortcomings addressed. So it remains a desktop, ARM based, co-operative multitasking OS. Fix what's broken, add what's missing but don't change what already works (even though this may conflict with some peoples belief systems about an OS, COS vs POS being the biggest!). Once that has happened then you can contemplate changing how the OS works.

mripley on 03/08/06 08:26AM [ Reply | Permalink | Report ]

Gunnlaugur has a good point, if RO were open sourced you would probably find it being taken in a dozen different directions, and end up with a dozen half finished versions, because there is no agreed roadmap for its future development either amoungst users or the programmers that might work on it. Everyone is concerned addressing what they percieve is a particular fundemental weekness that has to be tackled above all else, and no one is considering how to re-enforce the few things that give RISC OS its unique selling point.

As for embedded work, this can filter back many useful things to desktop RISC OS. For example improved networking with DHCP support, STB browser development (although with the continuing delay of O3 you could be forgiven for overlooking that), plus PCI and USB support which were first developed for embedded devices. All sorts of new things could come out of future STB work, such as support for streaming video to graphics cards from DTV or DVD. Even things such as synchronising between PDAs and remote severs, or downloading music to MP3 players are features proposed for future STBs.

druck on 03/08/06 09:41AM [ Reply | Permalink | Report ]

Lots of things *might* filter back from embedded work, but that's assuming there's a viable desktop operating system still available and that there's the motivation (financial or otherwise) to improve it with those new features.

I'm not saying that open sourcing RISC OS is a magic bullet. However, even if you end up with six different open source flavours, it surely has to be better (and possibly cheaper) than having two closed source flavours that arguably don't meet their users' expectations and that can only be improved behind closed doors by developers specifically hired for the task.

davidb on 03/08/06 12:13AM [ Reply | Permalink | Report ]

No, more than one open source branch would be a complete waste of time for everyone involved, with feature sets and backwards compatibility of each varient diverging away from the others (far more than between Select and RO5 where there was effort to maintain compatibility and APIs between the ROL and Pace variants throughout their development), no one in their right mind would use such a system. For any open source project to suceed it needs a well defined roadmap and tight control of the development process.

druck on 03/08/06 2:44PM [ Reply | Permalink | Report ]

I wasn't advocating six divergent branches of RISC OS, though I agree that it would be a waste of time for lots of people if that happened. A roadmap and tight control of the development process may well be beneficial, but they don't seem to be an obvious part of the current development model, either.

I just can't see the current players making RISC OS open source, anyway. Despite the advantages (both in terms of developer effort and user morale) that opening a central repository for the source code would have, it would probably take a leap of faith on the part of the current maintainer(s) and a lot of convincing before it could happen. Add to that all the feuding and in-fighting and it looks even less likely.

Maybe some corporate entity will eventually come along in the next five years and throw resources at the OS. Perhaps desktop users will eventually see these benefits. It gets less and less credible to persuade people to stick with the platform based on how well it's doing in embedded markets. Convincing developers to write desktop applications based on that is even more of a long shot - and applications are what people really need.

davidb on 03/08/06 3:10PM [ Reply | Permalink | Report ]

Interesting discussions here..

It appears that we've already have made the step to where there are many opensource RISC OS versions :-)

Just for the sake of it:

If there is to be an opensource version of RISC OS it will be very unlikely that there will be a many versions approach. There aren't enough programmers around to do such an enterprise..

If people would be more educated about opensource and looked at (for example) the linux kernel they will see that is offers the linux users the opportunity to choose pre-emptive and non-preemptive kernels (not user space though). There are special patches for Real-Time extentions an so on.. None of these patches break the rest of the linux distribution and only affect very specialized user groups. None of these patches or choises has led to splitting the Linux kernel into 7 or more different branches that evolved into different directions.

The thing is that we keep speculating on other people's speculations!

To get out of this circular reasoning is to collect all the wannahaves and must-haves, get enough people to commit themselves both(/either) in terms of time and money , and start with accomplishing what neither Castle nor RISC OS Ltd have accomplished. A modernized version of RISC OS which allows users to choose and do things he they want to be able on the platform.

There are still enough of us left to get this started properly and also to build the components needed. Doomsday scenarios can always be fostered when the userbase and enthousiasm for RISC OS has faded and all hope is lost for those very few left..

Let's cheer up, we have a userbase that is very much sympathetic towards RISC OS. It is unique in many ways and it is this user experience that we would like to expand into a more versatile way.

so keep up the good suggestions and we might just accomplish something here.

Jan Rinze.

JanRinze on 03/08/06 10:21PM [ Reply | Permalink | Report ]

It might still be a good idea to release RISC OS' sources under a licence, wich enforces a single branch development process with one central repository.

JGZimmerle on 03/08/06 10:44PM [ Reply | Permalink | Report ]

I reckon we will see RO5 sources become available as some sort of open-source. If you remember back to how RO5 came about, it was down to Pace engineers somehow getting the source code out of Pace months before hand.

Now that Castle have effectively got rid of all their hardware and software engineers I reckon the same thing is going to happen all over again.

Lets hope something good comes of it. RO6 anyone? ;-)

nx on 14/08/06 12:05AM [ 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