Drobe :: The archives
About Drobe | Contact | RSS | Twitter | Tech docs | Downloads | BBC Micro

How to port RISC OS 5 to the RiscPC

Published: 30th Jan 2007, 01:16:49 | Permalink | Printable

Possible but not a top priority for RISC OS Open

The RISC OS Open team have revealed the technical ins and outs of producing a RiscPC build of RISC OS 5 as well as uncovering the development history of the operating system. Ever since the launch of the ROOL scheme, there has been a high level of interest in a RiscPC port, and a lengthy document was published over the weekend outlining the work required to produce such a version of RISC OS 5.

The briefing also deflates conspiracy theories lurking around the whispered topic of how the OS was developed post-1998, and how the different Pace, RISCOS Ltd and Castle strands of RISC OS came about.

History
Ex-Pace engineer Ben Avison, who authored the document on the ROOL website, said the RiscPC-compatible source code for RISC OS 3.6 still exists in the vaults - however after that, things get slightly complicated. The OS kernel is split into three streams, the 'trunk', the 'Ursula' branch, and the HAL branch. Any of them can be used for a RiscPC port, but each has its own pros and cons. Ursula split from the main-line trunk after the release of RISC OS 3.71, in order to form an OS for the doomed Pheobe project, and was later used to seed RISCOS Ltd's stream of RISC OS.

When Acorn's yellow bombshell was canned, parts of Ursula were merged back into the trunk, and the option to run the OS in 32bit mode was later added. This become the Pace "homebrew RISC OS 4". The trunk then split again, according to Ben, forming the HAL branch - in which chipset support was moved out of the kernel, more optimisations from Ursula were brought in, and development was taken over by Tematic.

A hardware abstraction layer for the Acorn IOMD and VIDC chipset, as found in RiscPC-class machines, appears to exist, and it makes use of undocumentated aspects of the RISC OS 5 HAL system.

In his essay, Ben noted: "There still seems to be plenty of interest in building shared-source RISC OS for the Risc PC, but there is much confusion about what would be involved in doing so.

"First, understand that this is not top priority for ROOL - ROOL's policy is that it's more important to focus on releasing all the components needed to do an Iyonix ROM and disc build first.

"Only after this will the remaining components be vetted for release, although those components that relate to RiscPC hardware support are probably good candidates to be the next on the list."

See the link below for more information.

Links

Building RISC OS 5 for the RiscPC

Previous: Neil Raine's widow to run half marathon for charity
Next: Thunderbird 'demo' port released

Discussion

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

Even though I still use a RiscPC as a second computer, I am quite happy to keep it running ROS4. I honestly do not see the need for a version of ROS5 that runs on it. It would be better if the people who are going to take advantage of the shared source intiative devote their programming and engineering skills to moving the OS forward. Multicore processors appear to me to be the way silicon technology is moving, with the appearance of a dual-core XScale from Intel last year, wouldn't it be better if attention was focused on making RISC OS use similar and no doubt faster chips that will appear in the future.

 is a RISC OS UserJWCR on 30/1/07 9:45AM
[ Reply | Permalink | Report ]

The linked article contains some fascinating information. Some nice titbits about a threading module in the ROOL forum as well.

Assuming that there are no show stopping problems with the license or the release itself, this could be the start of a very exciting few years for RISC OS.

 is a RISC OS Userkillermike on 30/1/07 9:46AM
[ Reply | Permalink | Report ]

Just to stimulate the imagination of fellow dreamers, here is a page with some portable, mostly ARM-powered, Linux capable computers.

[link]

 is a RISC OS Userkillermike on 30/1/07 9:51AM
[ Reply | Permalink | Report ]

As far as the RiscPC is concerned I would much prefer new software such as Firefox even if with slightly reduced capabilities. Work on RISC OS would surely be better directed at enabling access to faster ARM chips?

 is a RISC OS UserAW on 30/1/07 7:09PM
[ Reply | Permalink | Report ]

Crumbs - why would anyone want RISC OS 5 on a Risc PC ? I'm constantly dragging sprites out of my RISC OS 5 Iyonix and onto my Risc PC so that I can use the select version of Paint. And using Edit on an Iyonix after you've got used to the select version feels like leaping back ten years; gosh, the strange way RISC OS 5's Edit still does a combined cut and paste in one move is not cute - just bloody annoying. Nope, I have no intention of installing a downgrade on my RISC PC, thanks. I'm with JWCR 100% on this one.

 is a RISC OS Usermartin on 30/1/07 10:36PM
[ Reply | Permalink | Report ]

martin:

Does anyone still use Edit? Other text editors, which allow this kind of thing to be configured, are available...

I can also remember the howls of protest when ROL changed the clipboard operation in Edit, with claims that the old Copy/Move system was far superior. It looks as if you can't win. :-)

Though I'm in agreement about not porting RISC OS 5 to the RiscPC.

 is a RISC OS Userstevef on 30/1/07 10:55PM
[ Reply | Permalink | Report ]

Porting RISC OS 5 to the RiscPC give people who still rely on their old computers (either because of specific software and/or podules, or who can't afford a more modern machine) means that these people will be able to get an up-to-date OS even if ROL go bangsplat. It opens the RiscPC OS market to competition. This can only be a good thing.

 is a RISC OS Userrjek on 30/1/07 11:00PM
[ Reply | Permalink | Report ]

I think a RiscPC port of RO5 would be a good thing, although I appreciate druck's argument that it's really only APIs that need to be replicated to resolve the problem of writing apps for RO4 and RO5. However, I'd like RO5 on the RiscPC because I'd like to see only one branch of RISC OS, so all efforts can be focused on it, and it gets 100% of the effort and not 50%. Personally, I'd don't care which branch of RISC OS 'wins', but I think ROOL certainly have the hearts and minds, Castle ultimately own RISC OS, and ROL have not really achieved much in quite some time.

So, ideally in my eyes, we'd have one version of RISC OS suppporting all modern hardware, and as Castle was making the Kinetic RiscPC up until a few years ago, I'd say it counts as modern. Well, as modern as any RISC OS hardware is.

Martin's argument that in many ways this would be a downgrade is true, I bought an A9Home as I thought that version of RISC OS would suit me better than RISC OS 5, the main reason being alpha-blended graphics. However, I realised the ridiculousness of that situation, that I had to choose different hardware to get the OS I wanted, even though Iyonix hardware would have suited me better (I'd have liked the TV card).

So as inconvenient as it is, I think getting RO5 onto all modern hardware is the way forward, the split should never have happened in the first place. We have the opportunity to reverse it, and although that's going to be tough on a lot of users, I think it is necessary in the long run.

 is a RISC OS Userthegman on 30/1/07 11:36PM
[ Reply | Permalink | Report ]

"It opens the RiscPC OS market to competition. This can only be a good thing."

Not really. The time wasted on porting an OS in many ways inferior to RO6 to a diminishing market of 13-year old machines should be going into developing new features and porting to new devices. I can't see ROL getting too worried about such a port.

Apart from offering a second option, please explain why I might want RO5 on my RPC. Wouldn't you be much better off just implementing particular modules to RO6?

 is a RISC OS Usertimephoenix on 30/1/07 11:51PM
[ Reply | Permalink | Report ]

"I think getting RO5 onto all modern hardware is the way forward."

What if RO6 is ported to the Iyonix? Either way one group of users is going to lose out on features to begin with.

The most sensible thing for ROL to do now is to release some of their code under a shared/open source license. Managed correctly, their development would be accelerated and it would allow new features to benefit both camps.

 is a RISC OS Usertimephoenix on 31/1/07 12:07AM
[ Reply | Permalink | Report ]

Reply to timephoenix, I agree that competition in the RISC OS market, i.e. for the *actual OS* is not a good thing, it's just too small, and even if it were larger, like the Mac, would you want say, Apple, the owners of the OS making a 'cut' of the OS, and some other company making another 'cut' of the OS for older hardware, with subtle differences to trip up developers? Of course not, what we need, in an ideal world, is one solid, stable, actively developed version of RISC OS.

 is a RISC OS Userthegman on 31/1/07 12:16AM
[ Reply | Permalink | Report ]

Josh and Martin are absolutely correct which is why this development has more to do with politics than anything else.....think about it!

 is a RISC OS Usermripley on 31/1/07 8:40AM
[ Reply | Permalink | Report ]

timephoenix: I don't know if you've noticed, but RO6 is a) buggy, b) not available, c) in a questionable state of development. Don't you think it's better to put your bets on the I' OS that it is more likely to receive development, bug fixes, etc? I've already suggested why you might want to have RO5 available to RiscPC users, as well as other people's comments such as more developers for ROOL, and legal free emulators (like RPCemu).

ROL will never release their source code. PM's likely already made that decision, and won't back down on it - it's the only asset they have. Giving it up would mean giving up. The most sensible thing for ROL to do is, err, to do *something*.

 is a RISC OS Userrjek on 31/1/07 9:04AM
[ Reply | Permalink | Report ]

I agree that it now looks as if the future is Open RISC OS 5, since in future it is likely to develop more quickly and keep up with advances in hardware more effectively that the RO6 stream, simply because there is more effort available for development. I would expect open RO5 to catch up with and overtake RO6 before too long for that reason. Getting it running on all current platforms is therefore an important part of re-unifying the OS. If RO5 is delevoped as we hope, and available on all current platforms then it will become the preferred option. Personally I own both RO5 and RO4.39, and unlike some find no advantage in the ROL stream, since I never use Paint (or any other graphics program - I can't draw)or Edit (I prefer Zap) and 4.39 can't run Fireworkz 32 which I use on the Iyonix. Both versions are equally good for photo processing (which is really a function of the aplications anyway). I don't use or need the multi-user facility. So I have a slight preference for the RO5 stream, but it is only slight. The real advantage is the likely development of this through ROOL in the future.

 is a RISC OS Usermrtd on 31/1/07 9:46AM
[ Reply | Permalink | Report ]

martin: A bit short sighted because if everyone was using RO5 then ROL may at last make select work on RO5 and you never know actually make the tools you need available on RO5 as part of that so you don't need to use two machines.....

Again as on the other thread if we are all on one OS there is more chance missing software will be made to work on the version everyone is using as there is a valid market on RO5 to sell the updated product to...

It may allow some people to get experience making the OS work on diffrent hardware and develop skills to a point where they may be able to go on and help porting to more intresting devices not everyone can start at the deep end on harware they don't know...

 is a RISC OS Usermrmac on 31/1/07 10:25AM
[ Reply | Permalink | Report ]

Hopefully with the existance of ROOL the Select / Adjust / RO6 benefits will start finding their way into it anyway.

 is a RISC OS UserSimonC on 31/1/07 11:07AM
[ Reply | Permalink | Report ]

Well, that is the hope! We need to resolve the split, and I would prefer to see it resolved in a way that leaves us with open source. That means RO5, and if that becomes the dominant strain, I'm sure that any Select/RO6 features that are important to people will be added. This will probably happen more quickly than Select/RO6 on Iyonix.

 is a RISC OS Usermrtd on 31/1/07 12:57PM
[ Reply | Permalink | Report ]

mrmac: "Again as on the other thread if we are all on one OS there is more chance missing software will be made to work on the version everyone is using as there is a valid market on RO5 to sell the updated product to... "

I ask again, what missing software? There is nothing that runs exclusively on RISC OS 5 other than things which specifically require Iyonix hardware. If there isn't a valid market for software now, everyone running RO5 will make not a slightest difference.

RISC OS is a modular operating system where practically every part of it can be enhanced via replacing modules and via hooking in software vectors. When RISC OS source is available the individual components can be enhanced and soft loaded over the top of whatever existing OS version is required to support the machines hardware.

Development effort needs to go primarily on writing new applications, and secondarily on enhancing the operating system. No time whatsoever should be given to this obsession of getting the same number in everyone's Task Manager info box - as that is much difference running the same version it will make to anyone in practice.

Ignore this advice and all the will be left of this platform is a few people pointlessly fiddling with the sources trying reinvent drivers for 13 year old machines, while all users of RISC OS have deserted it because there is useful nothing to run on the OS regardless of any enhancements.

 is a RISC OS Userdruck on 31/1/07 1:42PM
[ Reply | Permalink | Report ]

erm I am not talking about stuff that runs exclusively on RO5 we are talking about all the software that doesn't run on RO5 of which I belive from what others are saying there is quite a bit...

 is a RISC OS Usermrmac on 31/1/07 1:47PM
[ Reply | Permalink | Report ]

Quite simply ROOL's RISC OS allows us to resolve the split ourselves, without needing to rely on ROL ever crossing over. Moreover, given sufficient market pressure, ROL could re-invent themselves by dedicating to adding special features to a commercial, yet fully compatible version of RISC OS 5 able to run on all platforms from free emulator to Iyonix. It could mean everybody working together.

As it stands with ROL being the company that it is, one has to keep paying good money with no real assurance what and when something arrives. I think ROL is in over their head, with A9home users still eagerly awaiting a complete and stable release of Adjust/Select32, subscribers still awaiting Select 4, their chief developer gone and a lot of people disillusioned with them. I see no other viable route, but to seriously consider joining the ROOL initiative while they still can.

If running RISC OS 5 on my RiscPC means helping to consolidate the OS split, I'll do so and most software I use will probably keep working. I've one RiscPC left, which runs RO 4.02 since Select caused problems outweighing the benefits. Putting RO5 on it would, in such case, probably enhance it and give a good feeling that it's basically running the same version as elsewhere. I feel the platform has always been able to rely on Castle, they've always been the dependable and enterprising guys here. I just don't get the same feeling with ROL, which rather makes me feel uncertain and worried about the future of RISC OS.

 is a RISC OS UserhEgelia on 31/1/07 1:50PM
[ Reply | Permalink | Report ]

hEgelia: If Select which was remarkably compatible caused problems on your Risc PC, what do you think running a 32bit only operating system will do? No 26 bit applications will run at all.

Even if Adrian does a version of Aemulor for it, what is the point? You are then back to using the same 26bit modules, with the same 28MB restricted wimp slot, but at a fraction of the speed it did when running a 26bit OS, and still plenty of compatibility issues such as with podule drivers. And to put the final nail in the coffin NO VIEWFINDER.

Its an utterly and totally pointless, of less value than even RISC OS 6. If I gave you a copy of RISC OS 5 for Risc PC today, you'd have deleted it and be back to running 4.02 an hour later, and complaining like mad about all the things I've detailed above. I haven't got a copy today, but how many months of effort do you think we should waste to prove that point?

 is a RISC OS Userdruck on 31/1/07 2:03PM
[ Reply | Permalink | Report ]

druck: You're going to hate an Open RISC OS, people will be forever doing things of which you disapprove, and there's bugger all you can do to stop them.

Embrace it, people will work on what they find interesting and useful TO THEM. If there needs overlap yours, excellent, if not it's not really going to affect you anyway.

From the actual document on the ROOL website "it should still be possible to build the OS to run in 26-bit mode".

"but how many months of effort do you think we should waste to prove that point?"

It's not your time to declare wasted.

 is a RISC OS Userflibble on 31/1/07 2:28PM
[ Reply | Permalink | Report ]

Tell you what lets agree to disagree on this...

I really don't want to deal with ROL either as last year I emailed them about the middle of the year to purchase RO roms from them and must have sent 4 or 5 emails (form different accounts in case that was a problem) to get the price so I could order their product and to date I have had no response at all from ROL...

How do they expect to sell a product if they won't reply to a customer who wants to buy the OS.... And I challenge you to tell me how treating a potential new customer (someone returning to risc os) like this isn't going to put a nail in the proverbial coffin?

Also I am not stupid and on examening the options I buy the os then I can subscribe to "the Select scheme" to get more functionallity... Who in their right mind that is returning to the market is gonna pay ROL money for select with the track record they have had so far....

You say you want the market moved on well lets make it happen and give everyone a product the can get behind...

As I said at the top of this post I can see we will never agree on this so I'll end the conversation after this post but take note a few thousand A9's and Ionix's dosn't constitute a market for any software developer and if you want all the software to be available on the markets fastest and premium hardware your gonna have to find numbers from somewhere. And I suggest it will take a lot longer to make ROOL available on a totally new device and attract new customers than it will to get it working on an RPC which people already own...

Well that's it i'll say no more on the subject otherwise we will still be posting this time next year...

Thanks for reading

John

 is a RISC OS Usermrmac on 31/1/07 2:31PM
[ Reply | Permalink | Report ]

mrmac: While I'm happy to say we differ on this, and sympathise with your views on ROL, this issue is too important just to brush aside. I know that despite my advice someone is going to waste time on this forlorn venture, because people such as you have asked for it. Therefore anyone who wants RISC OS 5 on the Risc PC must state precisely what they expect it to be, in terms of functionality and compatibility with both existing software and 3rd party hardware.

Otherwise if fiddle is right and all we get is people tinkering with the source for their own bloody minded amusement, ROOL is going very quickly plummet the to depths of customer dissatisfaction enjoyed by ROL, and the platform will be sunk along with it.

 is a RISC OS Userdruck on 31/1/07 2:54PM
[ Reply | Permalink | Report ]

Druck: "If I gave you a copy of RISC OS 5 for RiscPC today, you'd have deleted it and be back to running 4.02 an hour later" No /I/ won't. I'll be using to test against a 32-bit only OS. I don't care that it won't run a ViewFinder, that my 'favorite' desktop utilities (and sillies) won't run. I'll sit there with a copy of Armalyser and make all those utilities and sillies 32-bit :-D .

 is a RISC OS Userphilipnet on 31/1/07 3:25PM
[ Reply | Permalink | Report ]

druck:

In my mind we need to forget about 26bit for the RPC version. We already have a 26bit OS and I am not suggesting we re-invent the same thing as this will still result in 2 versions of the OS and solve nothing.

Yes! A 32bit version will break old 26bit software the same as it did on the Ionix. Only by creating a 32bit os that is as close as possible in all versions do I see this as a good, sensible and viable way forward.

I am in total agreement with you if anyone is suggesting we re-build a 26bit version from the ROOL source for the RPC. It will only result in a complete waste of time as we only loose features and gain nothing in terms of making it easier to develop software and it won't create a market for more software to be 32bitted.

Lastly if the RPC, Ionix & I guess A9 as well can run the same 32bit version of the OS so software developers can now justify developement of a 32bit version to the new enlarged pot of customers who require a 32bit version of software. On top of this 32bitting of software may present the developer with a good opertunity to add a missing feature as they will be ammending the code anyway and with just a single version of the OS to code for it will likely make any development of new features easier (though of course that may also not always be the case in some types of software that would benefit or specifically need new hardware features in later machines).

I hope this better demonstrates where I am coming from and where I see this as a potential good thing for the market rather than a bad thing.

ATB

John

 is a RISC OS Usermrmac on 31/1/07 3:32PM
[ Reply | Permalink | Report ]

Once there is enough of the code and tools in place to build an Iyonix version of RO5, it might not be a great deal of work to make a RPC compatible version of RISCOS 5.

So, the actual benefits of a RPC version are debatable but they do exist and it might not be a very difficult thing to do. I've no doubt that /some/ RO5 26/32 bit neutral RMs and support utilities will find their way over to the RPC.

I agree that making a RPC build of RO5 wouldn't be worth a *huge* effort at the expense of other projects.

Unless someone can correct me, I suspect it is possible to make a version of RO5 for the RPC which would boot in either 26 or 32 bit mode depending on user settings.

Obviously, the only other way that RPC users are going to get any RO updates is through Select.

 is a RISC OS Userkillermike on 31/1/07 3:57PM
[ Reply | Permalink | Report ]

druck: "If Select which was remarkably compatible caused problems on your RiscPC, what do you think running a 32bit only operating system will do? No 26 bit applications will run at all."

The problems were related to 2 main things: I couldn't get my Select to work with my screen (X-filed) and one often used application would crash at a particular point probably due to restructured OS components in Select. It had nothing to with 26-bit / 32-bit compatibility. Like I previously said; most software I use will probably keep working - meaning they are 26/32-bit neutral and rely on common 4.0x / 5.xx aspects.

If you read the ROOL article referred to by above Drobe news, you'd find this bit:

"There is no provision for mixing 26-bit mode and 32-bit mode applications on the same machine – if you built the OS to run in 32-bit mode, you would be losing the ability to run legacy applications. However, it should still be possible to build the OS to run in 26-bit mode, with the loss of large application slots and the large module area."

Like flibble also pointed out, RISC OS 5 can be built to run in 26-bit mode and, furthermore, there's probably a fair deal of support left for the RiscPC in the sources available to ROOL. This was the point I was trying to make earlier; the effort required to make it RiscPC / emulator compatible may not be so substantial, though certain obstacles need to be overcome. The price of OS consolidation, I guess.

Yes, there are a load of 26-bit apps left, but clinging to those might also severely restrict future growth possibilities - personally, I do not require them just as I no longer require the name 'Acorn' to play a role in my RISC OS computing. I like to run a relatively modern OS with relatively modern apps on it, like NetSurf for example.

"If I gave you a copy of RISC OS 5 for RiscPC today, you'd have deleted it and be back to running 4.02 an hour later, and complaining like mad about all the things I've detailed above."

That depends - but if you cooked up a special version for me you may well be right ;)

"Therefore anyone who wants RISC OS 5 on the RiscPC must state precisely what they expect it to be, in terms of functionality and compatibility with both existing software and 3rd party hardware."

Fully compatible. I'm pretty sure most stuff will work like it did with 3.7, since the resources still exist + the various Pace / Tematic builds continued for the RiscPC until about 2001. I'm even willing to give up on certain things, if it means working together on a singular OS with all accompanying benefits for the market and developers.

"Otherwise if fiddle is right and all we get is people tinkering with the source for their own bloody minded amusement, ROOL is going very quickly plummet the to depths of customer dissatisfaction enjoyed by ROL, and the platform will be sunk along with it."

I'm not so sure about that or if ROOL would allow it. We still have to see the final license which will determine the specifics. I'm glad you care about it so much, but please keep an open mind.

 is a RISC OS UserhEgelia on 31/1/07 4:06PM
[ Reply | Permalink | Report ]

@hEgelia

You make another important point about a 26 bit version being emulator compatible. This opens up the possibility of bundling an emulator with ROM images as a free package.

 is a RISC OS Userkillermike on 31/1/07 4:21PM
[ Reply | Permalink | Report ]

I think another good argument for providing RO5 on the RiscPC is that this may attract 'hobbyist' programmers to contribute to RISC OS. It's one thing to convince a hobbyist to shell out maybe £100 on a RiscPC so he can play around with RO5, and maybe contribute some useful code, it's another to get him to shell out £800 on an Iyonix.

If we get a port to the RiscPC, and it attracts just a handful of enthusiastic hobbyists to contribute useful code into RISC OS, then it'll have been worth it. I'm talking about the kind of user who might otherwise contribute to Haiku, SkyOS, AROS, that sort of thing.

 is a RISC OS Userthegman on 31/1/07 5:37PM
[ Reply | Permalink | Report ]

thegman wrote: "It's one thing to convince a hobbyist to shell out maybe £100 on a RiscPC so he can play around with RO5, and maybe contribute some useful code, it's another to get him to shell out £800 on an Iyonix."

That's £100 for the RiscPC and £50 for the tools. ;-)

 is a RISC OS Userdavidb on 31/01/07 6:38PM
[ Reply | Permalink | Report ]

Reply to davidb, I stand corrected. :-)

 is a RISC OS Userthegman on 31/01/07 6:41PM
[ Reply | Permalink | Report ]

killermike: The 26/32-bit mode switch is a build-time switch. I wouldn't recommend making it into a run-time switch, because it would have to be checked on every entry to the kernel interrupt, SWI and exception handlers. But you could store both 26-bit and 32-bit ROM images on your boot disc and adapt the softload tool to give you a choice of which version to boot into.

 is a RISC OS Userbavison on 31/01/07 10:40PM
[ Reply | Permalink | Report ]

thegman wrote: "It is one thing to convince a hobbyist to shell out maybe £100 on a RiscPC so he can play around with RO5, and maybe contribute some useful code, it is another to get him to shell out £800 on an Iyonix."

Which opens up the age old question of - Why is modern entry level RISC OS equipment so expensive? At £500 even an A9Home is too expensive. The platform needs something equivalent to a £250-£300 machine with a degree of upgrade-ability that will get hobbists interested in our favourite operating system. As their interest develops, then they will move on to higher specced more expensive machines.

 is a RISC OS UserJWCR on 31/01/07 11:08PM
[ Reply | Permalink | Report ]

Reply to JWCR, I think the price of new RISC OS machine for the moment is unavoidable, the economies of scale simply mean they'll never be cheap. However, an open source RISC OS might mean it gets ported to cheaper hardware which was never intended to run RISC OS, devices like the Nokia N800.

I bought a new A9Home, but got rid of it as it felt like the full version of RISC OS would never appear, the price was an issue, but not the end of the world. I think user upgradeability is not that much of an issue, really the only thing the A9 lacks is room for extra hard disks, which can be be overcome with a NAS, lack of upgradeable RAM, which on RISC OS is not *that* big a deal, and lack of PCI, and what's actually going to go in those PCI slots? A TV card would be nice, a better graphics card is not really required without games to use it, so I think the A9 is actually a pretty sound machine, so long as it has a solid OS for it, which it does not.

 is a RISC OS Userthegman on 31/01/07 11:15PM
[ Reply | Permalink | Report ]

Quite right. In the Set-Top Box (STB) world, we were usually aiming to get the bill of materials cost (BOM) for your basic ARM-based RISC OS STB down to under $100 - that's under £50 in today's money. But that's with volumes of, say, 50000 machines. If you want to produce the same thing in volumes of, say, 50 machines, you will find the BOM cost is painfully higher. And that was tried-and-true ARM9 based stuff so if you want to use whizzy new CPUs, RAM and other bits you'll be paying even more for them when placing small orders.

All that means you're highly unlikely to see a native RISC OS hardware design that's cheap so a port of shared source RISC OS 5 to some commodity ARM-based hardware would be the low cost solution.

 is a RISC OS Userriscosopen on 31/01/07 11:29PM
[ Reply | Permalink | Report ]

"...a few thousand A9's and Iyonixes doesn't constitute a market for any software developer and if you want all the software to be available on the market's fastest and premium hardware your gonna have to find numbers from somewhere."

The question remains exactly how many users you would gain - although there is still a substantial amount of people using RPCs how many of them are going to be willing to give up their Select features, support for Kinetic, Viewfinder and possibly legacy applications? With the machine no longer in production and a diminishing second hand market, the numbers are going to get even smaller. And until you manage to get most RPC users to switch, the OS split will divide the market further: at this stage most 26-bit users are using RO4/3.7.

Exactly where all these exra developers are coming from also remains a mystery to me. I would find a RISC OS port to a modern Tablet or PDA is exciting, not a slightly different flavour of the same OS on a very outdated desktop. And don't forget in turn you've reduced the amount of developers available to work on 32-bitting legacy apps.

From a marketing perspective, such a move is just plain balmy.

Steve: Do you think the more powerful XScales will be utilized in the STB market to handle more interactive applications?

 is a RISC OS Usertimephoenix on 01/02/07 01:09AM
[ Reply | Permalink | Report ]

JWCR - whilst I agree RISC OS hardware is too expensive, a cheap product won't happen, or even help. Look at the A3010 - it was as cheap as an Amiga/Atari, cheaper than a PC, had the marketing of the BBC/C4/Acorn to help, was still used in schools, and it was faster than most Archimedes, it still didn't have any success.

As long as you can get a MacMini for 400ukp including VAT, which could also run Linux and Windows (and therefore VirtualRiscPC too) I really can't see [m]any non-enthusiasts buying an A9home for 500+VAT even if it could run RO5/6

An Iyonix2 would have to be considerably more powerful (i.e. at least break the GHz barrier, include AGP8x - although even that's dying technology now....) to be taken seriously by most too, especially since Marvell of all people own XScale - at least with an Intel badge it had some PR weight.

Moving away from the desktop (to tablet or sub-notebook) and more into STB/phones seems to be the best way for RISC OS (if we must stick to ARM) to go.

 is a RISC OS Usersimo on 01/02/07 12:31AM
[ Reply | Permalink | Report ]

@Simon John "Moving away from the desktop (to tablet or sub-notebook) and more into STB/phones seems to be the best way for RISC OS (if we must stick to ARM) to go."

I agree that niche platforms should be the goal for RISCOS development. I've often said that RISC OS is settings itself up for failure when it tries to be a 'Windows beater'.

I'd disagree about the A3010 being especially good value though. I seem to remember that it was about 100 pounds more expensive than the Amiga 1200. The PC might have seemed a bit more expensive but when you factor in the monitor and the harddrive and maybe 16bit audio IO and a NIC, the PC was probably less expensive overall thanks to cut throat mass market economics.

Speed wise, the ARM250 chipset was only a 50% improvement over the 1987 hardware.

 is a RISC OS Userkillermike on 01/02/07 1:21PM
[ Reply | Permalink | Report ]

Price price price......... That is the bottom line.

Take the xbox vs the PS2/3; a battle royal until the Wii came along. Why was the Wii successful? Well it was price. and a significantly cheaper price than a PC.

What is required IMHO for the future success (note, success) of RISC OS is to have it in a cheap box costing £70, or £100 with a harddrive. This means mass market infiltration and that means non-techie people. Techie people like us should put on our Joe-Public hat one of these days.

I beleive the ideal machine is an STB with freeview (yes the TV thing) and a remote that doubles as a keyboard/mouse for internet and email.

So someone is in their flat watching the telly. The programme is mince and they then decide to go online without changind rooms or turning on a computer.

If it cannot be done for less than £100 then there will be NO succes for RISC OS in the future.

The strength of RISC OS (IMHO) these days is not as a desktop computer but as a convenience device. TV/email/web (including skype).

Just my 2p worth. If you think that RISC OS has a successfull future as a mainstream desktop, then, with the greatest respect, you are wrong.

Regards Bob

 is a RISC OS Usernijinsky on 01/02/07 1:30PM
[ Reply | Permalink | Report ]

Simon wrote>"especially since Marvell of all people own XScale - at least with an Intel badge it had some PR weight."

Not completely true. As it happened Intel *held onto* their ARM Architecture license and still *do* produce new ARM based processors (the dual 1.2GHz core ARM for use in RAID array cards being one). Yep they're called IOP's rather than xScale but we know what they mean (I'd also point out Iyonix uses an IOP albeit the less whizzy 80321). Marvel as I understand it got the "other" xScale brands.

 is a RISC OS UserAMS on 01/02/07 1:33PM
[ Reply | Permalink | Report ]

Michael Reed: It's possible that the A3010 didn't sell in high street shops because the salespeople didn't get a big enough bonus for selling them and so never displayed them properly, never pointed them out to customers, and never learnt how to switch them on. But you are so right about the best options being in niche markets.

 is a RISC OS Userjc on 01/02/07 2:31PM
[ Reply | Permalink | Report ]

Michael Reed:

A3010 vs A1200 compared well I thought. Games conversions worked just as well on the Archie, despite the lack of graphics acceleration. However I think the A3010 made a better value desktop computer for serious work, despite the cooperative multitasking, as it was about 4 times faster, had a high density floppy drive, and had the whole desktop OS on ROM. The ARM250 may have been only 50% quicker than a 1st generation Archie, but it still compared well to the competition in 1992, especially in an entry-level machine. By 1995 the A3010 had dropped in price to below the 300 pound mark, but the Escom-relaunched A1200 still cost 399 (the same price as the original). Obviously you could get all sorts of CPU upgrades for the A1200 years later, but that's because it was never really superseded. In the Acorn market everyone who wanted a faster machine upgraded to the RiscPC.

As for 1992 vintage PCs I think you overestimate their economy. A budget 386 would have cost a thousand pounds and a Soundblaster would have been a luxury, much less a NIC. The A1200 and A3010 could just have plugged into a TV, no need for a monitor.

Anyway, this is ancient history and off-topic.

 is a RISC OS UserCogs on 01/02/07 4:05PM
[ Reply | Permalink | Report ]

nijinsky: "The strength of RISC OS (IMHO) these days is not as a desktop computer but as a convenience device. TV/email/web (including skype)."

I'd accept that there's a lot of potential for devices rather than the "... PC" where Bill Gates is frequently on hand to fill in the dots, visionary style, only to be proven wrong time and time again. But can RISC OS really compete in this sphere? The hardware constraints for things like Web-on-TV boxes, PVRs, personal e-mailers and so on aren't nearly as harsh as for things like high volume mobile telephones (where RISC OS supposedly has a big advantage with regard to memory usage, but would probably need a lot of work to make it robust enough, thus eliminating it from that sector), and the flexibility of solutions like Linux seems to make them somewhat more attractive for device makers, even if some of them don't quite understand the licensing obligations (which are really fairly mild unless you're a clueless "we must keep everything ultra top-secret!" corporate freeloader). Indeed, the ready availability of software for platforms like Linux arguably permits device makers to concentrate on differentiating or improving their hardware, and the task of getting Skype on that hardware (if that were deemed important or acceptable) would be more strategic than technical. Moreover, they would probably be able to look more closely at some of the human factors issues that seem to get otherwise ignored when pushing stuff out to market.

In any case, if RISC OS does have a future in the device market, what we may not see (in contrast to Linux) is the continuity in the community where people doing desktop stuff might consider turning their attentions to embedded stuff, and vice versa, leveraging what they already know in the field which is unfamiliar to them. RISC OS is already quite a fragmented thing, and I can't see the embedded and desktop branches constituting a single meaningful platform over the longer term. And perhaps the fate of BeOS has lessons for us there as well.

 is a RISC OS Userguestx on 01/02/07 5:08PM
[ Reply | Permalink | Report ]

GuestX: "RISC OS is already quite a fragmented thing" but it isn't, there is just an irrational obsession with only having one OS version. The most important issue is the overwhelming vast majority of software both old, current and new runs on all versions of RISC OS which means it is not fragmented, the APIs remain compatible so it doesn't really matter how many development strands there are apart from the obvious issue with duplication of effort amongst limited programming resources.

There are dozens of Linux distributions, and you don't see the Linux community in paralysed by the prospect and spending all their time debating how to get back to the "one true distro". As long as the effort continues to standardise the APIs, each distro can target its specific strengths and areas of interest.

 is a RISC OS Userdruck on 02/02/07 09:04AM
[ Reply | Permalink | Report ]

druck: "but it isn't"

There are two different "owners" (more like licensees as far as we know) scrambling around on different branches of a proprietary codebase, developing their own APIs, playing leapfrogging version numbers, with people running different vintages of both of them (although probably a lot more running different versions spawned by RISC OS Ltd., due to people bailing out when they didn't get their updates), and that isn't fragmented? At least with Linux there's a thorough understanding of how all the parts fit together, you can freely combine them, limited only by compatibility issues between tightly bound components. And here's the crucial thing: on Linux you almost always have some obvious migration options if you start to disagree with where your distro is going, whereas the RISC OS scene seems to rely on continual good luck that people will still be able to open their applications to get at their data if they move over to another flavour of the operating system. And people wonder why Select never came out for the Iyonix, let alone why RISC OS 5 (or perhaps they'll call it RISC OS 7) will ever make it onto the Risc PC.

 is a RISC OS Userguestx on 02/02/07 11:50AM
[ Reply | Permalink | Report ]

GuestX: two words - absolute rot.

 is a RISC OS Userdruck on 02/02/07 1:22PM
[ Reply | Permalink | Report ]

druck: "absolute rot"

What? The bit about having the freedom on other platforms to choose where you go with your data, or the bit about RISC OS's multiple "owners", issues with getting binaries to work on newer versions of the operating system, and the fact that many people are actually confronted with a mess caused by binary-only abandonware?

 is a RISC OS Userguestx on 02/02/07 2:03PM
[ Reply | Permalink | Report ]

Reply to druck, All due respect, but I'd say the Linux distribution method *is* a total mess, with regard to countless different distributions all distributing their own way of installing software, i.e. apt-get, portage, yum, rpm, that one Linspire has, and god knows what else. Also you've got the various toolkits which ensure no app looks/behaves like another.

On the good side, like guestx says, at least it's all there for the taking, and nobody is stopping you from using one part of a distro on another, and it's a big enough market to support it, and there is enough commercial interest from BIG companies who don't mind supplying their goods in a few different formats. They also have the manpower to test on various distros.

You could argue that the proliferation of distros on Linux is a good thing, maybe it keeps people on their toes, healthy competition etc. But I really don't think RISC OS can stand that, if we had perfectly compatible APIs, then maybe, but of course we don't. Do we even have fully compatible Standard C Libraries on all platforms yet? We could work towards getting rid of all the API compatibilities and say to ROL and ROOL; 'OK guys, play nice, make your stuff work the same' and /hope/ that both sides do just that. Or we can say 'Look, ROL, you had your chance, you did a good job in the beginning, but now it's just not working out. We're all going to support one version of RISC OS so we're not always re-inventing the wheel.'

 is a RISC OS Userthegman on 02/02/07 5:09PM
[ Reply | Permalink | Report ]

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

Search the archives

Today's featured article

  • Running a fashion label with RISC OS
    Independent t-shirt design in Germany
     3 comments, latest by Sawadee on 22/10/05 9:36AM. Published: 16 Oct 2005

  • Random article

  • ARMDebug update: Singlestepping, OS trapping, memory editing wonder

     Discuss this. Published: 15 Dec 2000

  • Useful links

    News and media:
    IconbarMyRISCOSArcSiteRISCOScodeANSC.S.A.AnnounceArchiveQercusRiscWorldDrag'n'DropGAG-News

    Top developers:
    RISCOS LtdRISC OS OpenMW SoftwareR-CompAdvantage SixVirtualAcorn

    Dealers:
    CJE MicrosAPDLCastlea4X-AmpleLiquid SiliconWebmonster

    Usergroups:
    WROCCRONENKACCIRUGSASAUGROUGOLRONWUGMUGWAUGGAGRISCOS.be

    Useful:
    RISCOS.org.ukRISCOS.orgRISCOS.infoFilebaseChris Why's Acorn/RISC OS collectionNetSurf

    Non-RISC OS:
    The RegisterThe InquirerApple InsiderBBC NewsSky NewsGoogle Newsxkcddiodesign


    © 1999-2009 The Drobe Team. Some rights reserved, click here for more information
    Powered by MiniDrobeCMS, based on J4U | Statistics
    ""
    Page generated in 0.3815 seconds.