Five tips for ROL over the next five years

By Paul Stewart. Published: 10th May 2009, 20:38:49 | Permalink | Printable

Opinion - Drobe man Paul Stewart offers a few suggestions to RISCOS Ltd to ensure the company and its platform survives over the coming years.

I believe that RISCOS Ltd has come a long way since its formation in 1999, and has developed RISC OS 6 into a very good product. ROL eventually reworked its branch of the OS to be 32bit neutral, allowing it to be built for newer ARM-compatible processors, and persevered with its RISC OS Select subscription scheme to push OS updates out to paying end users.

That was then. Now we're well into 2009. So what's required from ROL to see us through the next few months and years? I have prepared a few ideas for the company's management to consider, and also for Drobe readers to debate.

1. A roadmap
Wouldn't it be nice if we could be shown a generalised outline stating where ROL expects to be with RISC OS in five years' time, with Select releases and other projects marked on it, plus yearly goals detailing how RISC OS will be improved? This could at the very least let Select subscribers know what to expect and could even entice more users to the scheme. Publishing this roadmap document would also counter those who believe a lack of vision is one of RISCOS Ltd's major failings.

2. Transparency
When RISCOS Ltd was a fledgling company, it produced a report at the end of each year detailing in general how the company was progressing. Whatever happened to this? Why was it stopped? How about reviving this practice so the community can see how the company is doing? When the Select scheme was founded, punters could contribute to a wish list - a database of features that subscribers wanted to see in RISC OS. Does this get used when deciding where to go forward and where not to? Does RISCOS Ltd care about what the users want, or is it just what the developers want to do?

Other than the RISC OS Open initiative, ROL has no direct competitors anymore, and it keeps insisted itself and ROOL are cooperating on development, so there's no commercial advantage to be lost by sharing its plans with the close-knit community surrounding it.

3. Involvement
In such a small market there are obvious financial constraints that prevent the employment of customer support staff, sales representatives, marketing teams, project managers and full-time teams of software developers. Yet, if someone has a lot of free time and is willing to get involved with ROL's projects there appears to be no open way of doing so - even though ROL says third-party developers are welcome to contribute, the project feels like a closed shop. Compare this to RISC OS Open, which actively relies on user participation.

I would hope ROL appreciates any offers of help from programmers and other suitably-skilled users who can offer a few hours a week of their time, but there is nothing on the website that says this is even considered. I know the business model of RISCOS Ltd is quite different to that of RISC OS Open, but I would still have thought this type of involvement would be welcome.

4. Web presence
It's nice to see after 10 years the ROL site is now starting to get some attention and has dropped its original jigsaw-piece design. RISCOS Ltd should look at other websites, such as Apple's Mac OS X site, for inspiration. If more information was in the public domain about involvement, perhaps this website have been updated a while ago. Not only did the previous design looked dated for some time, there are still out-of-date documents, such as the mission statement and suppliers list. This is certainly not a good advertisement for the company or RISC OS in general.

Although it's taken almost 10 years to discover the power of the internet, finally RISCOS Ltd has opened an sales website. From here both Virtually Free RISC OS and the RISC OS Upgrade CD can both be purchased. What about Select subscriptions and the RISC OS 6 PRM? Why are these not on the sales site and available for punters to buy? These obviously were not in the train of thought when creating the sales site, however I see no reason for them not being there. Perhaps a little extra development work on the sales site could enable this.

5. Select developments
Overall, if we ignore the long gap between Select 3 and Select 4, the Select scheme has been a good success. Both RiscPC and Virtual RiscPC users have enjoyed regular updates to the OS and its ancillary applications. The question is how far should such a scheme go?

I think RISCOS Ltd has missed a trick here - after the fuss over ROL producing a version of Select for the Iyonix, the Select scheme should progress from being all about the operating system to being a product for the platform as a whole.

Would it not be feasible for Select development to be split, say six months on operating system work and six months on applications for the platform - such as finishing off the port of Firefox, porting the Gnash Flash player or an open source Java engine?

These are only examples. I’m sure other people would have their own ideas for general developments that would benefit the platform as whole. Such activity would almost certainly make users of RISC OS 4, 5 and 6 consider subscribing to the scheme and thus channelling additional funds into new developments. Sadly, it's unclear just how much in the way of extra funds this may bring or how effective this split development approach will be - but I'm sure it's worth considering.

Conclusion
Despite the criticism levelled at RISCOS Ltd, rightly or wrongly, I believe the company is certainly relevant in the RISC OS world of today. It is the only company that has constantly developed the operating system over the last decade. It has added extra networking features, improved the desktop experience, updated Paint, Draw and Filer and carried out numerous under-the-hood improvements. It is open to work with partners who wish to exploit the features of RISC OS commercially. Don't forget, RISCOS Ltd is not a large company. It operates on a small budget using a handful of part-time programmers to achieve its goals.

RISCOS Ltd should be about ensuring the continued development and availability of RISC OS, whether it is run on real hardware or via emulation, and the above list should go some way to making sure work is seen to be done and that work is being done to our favourite operating system.

Links

RISCOS Ltd website

Previous: Entry-level animation app Flicker goes on sale
Next: First screenshot: Beagleboard runs RISC OS 5 desktop

Discussion

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

Your enthusiasm is fantastic, Paul. And thanks for another interesting article.

You're right, RISCOS Ltd is a tiny company. But it could make a massive contribution to the platform. It should open source its branch of RISC OS so that the community (which is also small) can make the best use of the development it has done over years (which I value highly) in combination with RISC OS Open.

The subscription scheme could continue, allowing enthusiasts to fund important and democratically decided development. The code produced would be open source also.

While RISCOS Ltd remains closed, access to the small amount of money in my bank account are closed also. I'm no longer interested in funding cul-de-sac development.

fylfot on 10/5/09 9:22PM
[ Reply | Permalink | Report ]

I agree; it's difficult to think there's even enough money left in the market to pay Paul's salary (if he even draws one). The continued life of RISC OS is not directly connected to the continued life and feasibility of a company. I entirely agree that RISC OS's future is in a more open, community-driven system, rather than ROL's current closed, mysterious, and apparently arbitrary system.

rjek on 10/5/09 9:51PM
[ Reply | Permalink | Report ]

Open source is not necessarily a magic bullet which solves all issues and leads to development.

I think a clear direction is the most important thing they need.

markee174 on 10/5/09 10:42PM
[ Reply | Permalink | Report ]

Nobody suggested it is. But just look at the developments the ROOL branch has had since anybody can easily dip into it. Features people actually want have been implemented! It's clearly working for them.

All we need now is ROL to release their sources under a sensible licence, and CTL to change the licence ROOL are forced to use for something more reasonable.

rjek on 10/5/09 11:55PM
[ Reply | Permalink | Report ]

How much is the CTL licence a real stumbling block? I can appreciate it's not ideal - AIUI, it's not properly open-source, and contributions are something of a hostage to fortune if CTL changes its mind or goes under. But given the commercial realities of the RISC OS community, I wonder if expecting the licence situation to change is hoping for too much? It seems to be functioning pretty well at the moment, despite the (no doubt well-founded) worries of potential developers used to a less restrictive environment.

lym on 11/5/09 11:18AM
[ Reply | Permalink | Report ]

lym: Stuck on ARMs, nobody to contact about actual pricing for commercial use, no real "openness" or ability to cope if CTL vanish and RO's new owners have a change of heart, etc etc etc. It's much too complex, abitrary, unclear, and badly-written.

rjek on 11/5/09 11:52AM
[ Reply | Permalink | Report ]

Fair points (and it is crazy that CTL seems so coy about commercial licences), but given the amount of money in the community (or not), I don't think we'll be seeing lawyers drafted in to alter it any time soon. I wonder if this is the best we're going to get?

lym on 11/5/09 12:23PM
[ Reply | Permalink | Report ]

It's not (just) the clear direction, so much as publishing at least an outline/roadmap so that we can see it. ROL may well have a clear direction, but it's not much use to us if we don't know it.

As someone who works professionally on a very closed OS, as well as uses the very open Linux OS, there are advantages and disadvantages to both: open source is not the panacea it sometimes appears to be, but OTOH it's not all bad either, just so long as there is a specific direction.

DS1 on 10/5/09 11:54PM
[ Reply | Permalink | Report ]

ROL need to just throw the towel in, give all sources to ROOL and then we can have one true version to run on all the possible new hardware that risc os can run on thanks to ROOL. My Risc PC has been in the loft for about 5 years and my Iyonix hasn't been switched on for about 2 years but if a netbook style device came out that RiscOS could run on i'd buy it. Anyone wanna buy my Iyonix lol

ianscottiyo on 11/5/09 11:09AM
[ Reply | Permalink | Report ]

I suspect lots of people want to buy your Iyonix. Ebay it. You'L=ll get a good proce I don't doubt.

Monty on 11/5/09 12:01PM
[ Reply | Permalink | Report ]

I'd like to see ROL do well, as in many ways I think their product is more polished than ROOL's version. But I fear that their famously bad marketing and prickly attitude to criticism has put a lot of people off. ROOL (and especially Jeffrey Lee's OMAP port) have generated a buzz. That's the most important thing. No one seriously believes that either branch of RISC OS has all the features most users need or want. But at least there's a sense that ROOL has a plan. If ROL wants to compete effectively, I think they need to articulate a vision for their product and get people enthusiastic about it. I'm sure this is possible (open source is not for everyone), but it'll take a lot of work.

lym on 11/5/09 11:27AM
[ Reply | Permalink | Report ]

Indeed, RISCOS Ltd has made substantial progress both visible and invisible, but I believe nobody can argue that they've behaved rather irresponsibly at times. In this regard most attention seems to go towards the Select for Iyonix mess, but I think their worst behaviour appeared in the post RO4.39/Adjust period. Particularly when selling useless subscriptions to Risc PC users, when it actually financed development of 32-bit RISC OS Adjust. I think they've actually ensured the future of their company over the backs of unwitting Risc PC users due to their closed PR attitude. Speaking of which, their public image and marketing just suck and it doesn't show promise of change.

It may well be that RISC OS 6 is superior in most respects, but ironically it's not even available for the A9home, which I believe was the catalyst for its inception. In fact, up until now it only runs natively on ancient hardware. Which is why it will fail to attract the use ROL needs to survive.

As has been published on Drobe, RISC OS 5 is now running on the latest generation of ARM technology. The door to the forthcoming line of ARM based Netbook hardware has been opened. It's too bad for ROL, but this counts pretty heavily.

I know I could run RISC OS 6 in the commercial Virtual RiscPC emulator on a modern laptop, but why the hell would I want to run the arguably best version of RISC OS on an emulation of ancient hardware on top of Windows or Mac OS X? No, it's just cooler and possibly cheaper to be able to run RISC OS 5 directly on modern hardware, no strings attached..

In the end, the shared source nature of RISC OS 5 made it possible to get it to run on the Beagleboard and, by extension, opens it up to Netbook hardware. This just might give it the edge over ROL's plans for RISC OS 6.

hEgelia on 11/5/09 2:09PM
[ Reply | Permalink | Report ]

Remember that there still hasn't been a non-beta release of the 32 bit RISC OS Adjust that all those RiscPC owners paid for. As for which is the most advanced, well, that's debatable. Most of the new features in ROL's strand are never used.

rjek on 11/5/09 4:37PM
[ Reply | Permalink | Report ]

"Most of the new features in ROL's strand are never used."

Care to quantify such a sweeping statement?

sa110 on 11/5/09 8:12PM
[ Reply | Permalink | Report ]

Sure. Just look at all the effort put into new-fangled APIs that no software uses because they don't want to be tied to a specific version of RISC OS. Look at all the ripping out of stuff from the kernel into separate modules that from the user's perspective just makes *mo. confusing.

Here are a few others that I believe next to nobody uses; * ZeroConf, * Firewall and aliased interfaces, * PCX, PNM, XBM, SunRaster import in Paint, * Additional keycodes for multimedia keyboards, * Built-in command line zip tools, * Hardware profiles, * Multiple user accounts, * mDNS.

The list goes on. I'm sure somebody outthere uses those hideous "curved" buttons and graduations to make an elegant and understand GUI as blingy as a souped-up Nova. However, the features that make RISC OS 5 different from ROL's offering are the ones that are most appealing;

* A HAL that appears to be easy to work with, * Unicode, * Built-in USB, * PCI support.

That enough?

rjek on 11/5/09 8:27PM
[ Reply | Permalink | Report ]

"...as blingy as a souped-up Nova"

New keyboard now required. Will collect when I next see you at a show. Ta.

adh1003 on 13/5/09 10:09AM
[ Reply | Permalink | Report ]

"Just look at all the effort put into new-fangled APIs that no software uses because they don't want to be tied to a specific version of RISC OS."

Well I know of at least one application that manages to work on all versions of RISC OS and users features of Select when running on a Select version of RISC OS. So it's not impossible to do.

64k colours (another select feature that no doubt you don't think anyone uses) is supported by none ROL supplied applications, those same applications also working on none Select versions of RISC OS.

sa110 on 13/5/09 2:43PM
[ Reply | Permalink | Report ]

Nobody said it wasn't impossible. It's just a lot of effort for little gain. (Few of the new APIs are actually *useful*.)

Well done for finding a feature that a handful of people use; the 64k colour mode support hack, which is only useful because ROL haven't made a release suitable for machines with a design from the last decade and a half.

I see you don't content the dozens of other features that no body uses. I still stand by that most of the new features ROL have introduced see little, if any, use, and the tiny handful that ROOL's branch has are all extremely useful and well-regarded.

rjek on 13/5/09 3:28PM
[ Reply | Permalink | Report ]

I'd personally like to see a freely available, softloadable, update to SpriteExtend that will cope with alpha-transparency. If it could also support the standrard RGBA format then that would be brilliant too. ROL shot themselves in the foot here by not releasing their update to the wider masses and make it usable for developers (even if !Paint etc couldn't edit them, applications could use them.) The current solution here (Tinct) is really not acceptable (as you know, it was originally written purely for NetSurf and has to contain horrible hacks for printing.) Also, by not being centralised it stops hardware accelleration, and doesn't support anything other than scaling -- CSS3 image-orientation anybody?

not_ginger_matt on 1/6/09 1:17PM
[ Reply | Permalink | Report ]

"ROL shot themselves in the foot here by not releasing their update to the wider masses..."

I think your comment about ROL shooting themseles in the foot sums ROL's entire strategy. Trying to sell updates to an OS to an ever dwindling user base will evetually fail, as they won't be able to support this strategy ongoing.

Thats why I believe ROL is insignificant, and why all users should stop further dividing the OS base by propping up ROL - Time to get behind ROOL, and only ROOL.

nx on 1/6/09 5:00PM
[ Reply | Permalink | Report ]

I disagree with you somwehat here. I was given a copy of RISC OS 6 to update Tinct to work with 64k modes, and I was fairly impressed with the amount their desktop has changed for the better since RISC OS 4. Simple user-facing things like DHCP and copy/paste throughout the GUI made life so much easier. The image thumbnailing was also good, and !Paint and friends felt nicer too. As a developer though, I consider the fact they haven't passed-back their non-user facing changes as both disapointing and harmful. The obvious examples being the image conversion module and alpha-transparency sprites. I'd also have liked complete documentation of the (thousands of?) bugs they've squashed so I could be aware as a developer of what to be careful of -- just a simple online database of modules and the changes would have done. I assume (danger, Will Robinson!) this is also true for the ROOL developments. All in all, I'd say that ROL didn't manage to get developers on their side enough, whereas ROOL are thankfully doing the opposite. The general issue with giving things away for free though is that Joe Public who can't contribute code and wants to have rounded buttons and a graduated background won't ever get to see them. Given the choice, I'd personally rather have a Unicode FontManager -- but I doubt a very high percentage of users even know the difference RISC OS 5's makes. I think you're right that ROL is effectively over, but let's not pretend they didn't achieve anything worth keeping. I hope they can work something out with ROOL and not have their work become 'insignificant' by passing across some of their advancements so they can be merged rather than re-implemented.

not_ginger_matt on 1/6/09 7:01PM
[ Reply | Permalink | Report ]

I agree with the items identified as being good aditions to RISC OS and as has been stated the simple changes to the UI for setting up networking makes things look and feel a lot slicker. As you also say ROL are between a rock and a hard place in that they need money to do the updates and hence have the Select scheme but because it is a payable option then not everyone gets access to the good additions like the alpha stuff. In an ideal work then ROL would give away for free the components that need wider adoption and then the whole user base could make use of them but I don't think ROL could survive in their present form if they did that however by virtually giving away RO4.02/RO4.39 as softloads to RO3.7/RO4.02 respectively then I at least think they are trying to extend the user base for their components

bluenose on 1/6/09 7:33PM
[ Reply | Permalink | Report ]

Given the choice, I'd personally rather have a Unicode FontManager -- but I doubt a very high percentage of users even know the difference RISC OS 5's makes.

So how much difference does the RISC OS 5 Unicode Font Manager make?

sa110 on 25/8/09 2:13PM
[ Reply | Permalink | Report ]

The difference between a page full of black squares and showing the full range of characters as intended.

druck on 26/8/09 7:35AM
[ Reply | Permalink | Report ]

Can't say I've come across a page full of black squares yet.

sa110 on 26/8/09 11:46AM
[ Reply | Permalink | Report ]

"I'd personally like to see a freely available, softloadable, update to SpriteExtend that will cope with alpha-transparency"

Have you tried asking Paul Middleton about this? It maybe something they could do.

sa110 on 1/6/09 8:22PM
[ Reply | Permalink | Report ]

People have begged ROL to release certain APIs and improvements to make them available for all (and thus more widely used) on numerous attempts. I think the only thing that ever happened was their SharedCLibrary. Acorn used to routinely do this sort of thing.

rjek on 2/6/09 12:02AM
[ Reply | Permalink | Report ]

Before we get into specifics about freely available softloadable updates for people who have never contributed to the Select scheme, perhaps we should ask those people who have contributed whether they think the contribution they have made to RISC OS development over the past ten years should be given away for free?

So if you are reading this discussion and are a current Select subscriber then please let us have your thoughts. PS. Please quote your Select Number to verify that you are a subscriber.

It seems that lots of you who are pushing for RISC OS to be freely available are those who have never paid for anything from RISCOS Ltd. Do you never pay for software, or is it just RISC OS that you don't want to pay for?

If I was a loyal Select subscriber then I would be questioning whether I should support RISCOS Ltd any more if they suddenly started giving away everything for free. I appreciate that some people don't use Select features in their applications because they choose to assume the lowest common denominator of Operating System in use rather than testing for the availability of features, and then gracefully telling the End User that a required feature is not available on their specific OS. This is one of the problems we have been discussing with ROOL, as too many applications simply check for the level of OS present or module version number, rather than explicit modules being present, with known capabilities. However if no-one uses the new features available in RISC OS Six for their own applications, then we will have a stagnant OS, where no-one puts in new features, because they want to make sure that a RISC OS 3.7 user can still use the application they have written. We obviously want developers to use the new features that are available, and the RISC OS Six web site details much of that information. We are not hiding it.

We have also discussed the possibility of certain RISC OS Six features being made available in 32 bit format, for use on the Iyonix, but other than for completely altruistic reasons, why should we give away our developments for free? If it had been commercially and technically viable to release Select 32 for the Iyonix, we would have done so. So why should we do so now for free?

riscosboss on 2/6/09 12:53PM
[ Reply | Permalink | Report ]

"Before we get into specifics about freely available softloadable updates for people who have never contributed to the Select scheme, perhaps we should ask those people who have contributed whether they think the contribution they have made to RISC OS development over the past ten years should be given away for free?"

That would include me, then.

"So if you are reading this discussion and are a current Select subscriber then please let us have your thoughts. PS. Please quote your Select Number to verify that you are a subscriber."

Oh. That wouldn't include me, then.

Shame about those moving goalposts.

VinceH on 2/6/09 1:11PM
[ Reply | Permalink | Report ]

If I were a Select subscriber, and I was for most of it, I'd be more likely to ask why Adjust releases are so cheap compared to the years to funding I put in as a subscriber, and what happened to 3 years subs when there were no releases and coincidentally the A9 was in development? But lets forget about that, it's water under the bridge

What I certainly *would* *not* be complaining about is if 3rd party applications were able to take advantage of Selects enhanced features. Say for example Photodesk was able to load a greater range of image formats, and save sprites with alpha transparency, because the ImageRender and SpriteOP APIs were universally available.

Remember we are talking about low level APIs which allow RISC OS programs to utilise your version of RISC OS, not the user facing things such as fancy buttons, configure tools and new versions of Draw and Paint, which would remain exclusive to Select.

At the end of the day and OS is no use on it's own, it is there to run applications, and it's application development we are in desperate need of. Any boost to that by making it worthwhile to developers to use the features you've worked long and hard on, benefits everyone using RISC OS, yourself included. Attempting to keep everything to the ever diminishing number of Select subscribers, only hastens the day development of your branch of RISC OS will become non-viable, and all work done so far will be wasted.

druck on 2/6/09 1:17PM
[ Reply | Permalink | Report ]

Release some of the new APIs for free (not user-facing features) can only improve the situation. Currently, hardly any developers use them because it limits their software to an even tinier market. So this money's wasted, unless there's reason for developers to use them. And the only way that's going to happen is to make them more widely available.

Releasing those APIs for other systems makes the money the subscribers have spent with you actually do something, rather than nothing, and so it's better value for them.

"It seems that lots of you who are pushing for RISC OS to be freely available are those who have never paid for anything from RISCOS Ltd. Do you never pay for software, or is it just RISC OS that you don't want to pay for?"

I pay for lots of software; when it's worthwhile. RISCOS Ltd.'s product is expensive, and doesn't run on any hardware I own.

rjek on 2/6/09 1:17PM
[ Reply | Permalink | Report ]

I'm a Select subscriber (774, I think (Paul Stewart)).

Going back to what not_ginger_matt was saying. I think making available some specific modules,that can be used without the rest of RISC OS 6, perhaps something such as SpriteExtend or the Firewall is fine. I would however NOT go as far as saying make the new versions of Paint and Draw available or the Image File Render for free too.

sa110 on 2/6/09 1:29PM
[ Reply | Permalink | Report ]

IFR is an API. Don't expect more applications to use it unless everybody can. I've already said I wouldn't expect user-facing features to be made available (such as the filer "improvements", applications, the firewall, and err, whatever else there might be), however, APIs need to be widespread to be used.

rjek on 2/6/09 1:36PM
[ Reply | Permalink | Report ]

Indeed.

Earlier, Paul Middleton said "I appreciate that some people don't use Select features in their applications because they choose to assume the lowest common denominator of Operating System in use rather than testing for the availability of features, and then gracefully telling the End User that a required feature is not available on their specific OS."

Speaking with my sadly rarely worn developer hat on (which does have a nice peak and would have prevented the nasty sunburn I have on my forehead) I'm not likely to ignore a Select feature in order to support a LCD OS version - it's because I can't test anything I might write to use that Select feature to make sure I haven't included any silly errors as a result of misreading the documentation or whatever.

If you want developers to support it, make it available so that they can. Simples.

VinceH on 2/6/09 1:59PM
[ Reply | Permalink | Report ]

"IFR is an API"

Sorry, getting confused! However I think yourself and others are right. All these kinds of low level stuff should be released and availabe to all the developers, to encourage their use.

sa110 on 2/6/09 7:56PM
[ Reply | Permalink | Report ]

Not only the developers, but all users, too. Certainly at least back to supporting 4.02 and CTL's OS offerings. Just letting the developers use it isn't enough; why bother writing the code that only a proportion can make use of?

rjek on 2/6/09 9:25PM
[ Reply | Permalink | Report ]

For me the argument for freely releasing components (in a similar way that Acorn did with the nested Wimp, SharedCLib etc) is that it allows software to be written in an easier manner by developers. To me, the Firewall is most certainly not in that category.

I'm curious why you feel the Image File Render module isn't suitable for outside release. Given that it's effectively pointless on the surface to anyone without the Select thumbnailing Filer, no-one's getting anything for free. Similarly, to an end-user SpriteExtend is useless without Select !Paint.

The beauty comes when application developers get their hands on them. Want to write an application to load and render a PNG file? Simple. Suddenly you don't need to use libpng then waste your time writing some custom alpha-transparency plot routines. Similarly existing applications can be quickly updated to support exotic image formats transparently and centrally.

So, to summarise, you receiving benefits from increased application development and additional functionality. Without a large enough user-base to justify development the work wouldn't be caried out without freely available code (as has bene proven by the implementation rates.) All you can really be perturbed about is that others are also receiving some of the benefits.

not_ginger_matt on 2/6/09 2:03PM
[ Reply | Permalink | Report ]

Actually 774 is quite incorrect, having looked at my Select membership renewal.

sa110 on 2/6/09 7:51PM
[ Reply | Permalink | Report ]

Think of things from a developer's point of view for a moment. There is almost no chance of any commercial gain from RISC OS, so whenever someone writes code they are effectively donating to the platform.

Now, I'd always assumed the central purpose of RISC OS Ltd. was to drive forwards the RISC OS platform. So, you're effectively calling the few developers who are left helping RISC OS Ltd's cause a bunch of scroungers.

Well done.

Now, onto your lowest common denominator argument -- let's take NetSurf for an example. NetSurf needed an alpha sprite plotter, which only Select had. Should a window pop-up for every web-page a non-Select user visits? Of course not. I had to waste large amounts of time writing Tinct so that all users could get the same experience. That time could have been spent in far more constructive areas.

As for releasing RISC OS Six features for the Iyonix for free, how about the time I wasted updating Tinct to run on the 64k screen modes of RISC OS Six so you could make some money?

RISC OS Ltd. simply doesn't have the funding or, quite frankly, the leadership to do anything but wither. By working with ROOL perhaps you'll get the contributors using shared APIs rather than moving away from the RISC OS Select / Six variants and taking the precious developers and applications with them.

not_ginger_matt on 2/6/09 2:47PM
[ Reply | Permalink | Report ]

As a Select subscriber, though not willing to put my number in a public forum though i can send it to riscosboss if he likes, since the start, I have mixed feelings about how to approach this. On one hand I have invested a lot of money in the scheme but then I have had the benefits over the years before a lot of other people so in a lot of ways I have had something out of my investment. On the other hand very few applications make use of some of the features so that has limited the scope for maximising my investment. One way out of this is for ROL to make say the last stable set of API and non user facing features available for all i.e say take these from 4.39 and when RO6 goes to the next level, say 6.xx then make 6.xx elements available this means people can still get something out of the investment and the OS does move forward.

There are always people who want things for "free" and it's not just in computing but I do think moving to a rolling set of freely available updates makes sense though again only those at ROL can know if it makes financial sense.

What would make my investment even more value would be some Iyonix components and though late it may be better then never as they say but just a pity that it was only at Wakefield we were told that Select for iyonix was dead so i guess I would have to reserve judgement on that value to I actual see it.

Finally I would like to see ROL/ROOL show some public display of unity on moving forward if for no other reason that to at least silence the constant ownership debate though I don't hold up much hope given some of the entrenched positions on both supporters sides, though perhaps not from those involved in both ROL/ROOL.

bluenose on 2/6/09 7:01PM
[ Reply | Permalink | Report ]

Probably, I am too late to reply to this post, but I think it is worth anyway.

Mr. P.M., to me it sounds unbelievable that a business man like you has posted something like this, honestly.

I just hope this is a misunderstanding, as I imagine you should PROTECT the investment made by the Select-Scheme's Subscrivers and not waste their money on OS's features that the majority of Software Applications are not going to use.

I suggest you to release as soon as possible the low-level API to developers (let's be clear on this, I am NOT talking about apps like Paint or what is generally defined as complete software application).

A Select-Scheme's Subscriver for sure would like to see all the available software applications works better on his ROS Platform and see them to use all the OS's features. So, releasing the low-level API to every developer and every branches of RISC OS, is, probably, the best way to protect the S.S. Subscrivers' investment and wishes.

Moreover, from your own point of view, this is really good choice too, because improving the quality of available applications for RISC OS means to improve the value of YOUR work, not just in terms of fame, but also in economical terms, you know exactly what I am talking about.

Since the end of the "80's style computing" and the rise up of IBM PC's world and the Microsoft's business philosophy everybody knows that the value of a platform it's given by the software that runs on it and not by the UNUSED OS FEATURES. Also Linux, and the Open Source world, is a "victim" of this business view, as Linux is evolving mostly by the enterprise's side than the Desktop users' side. This is due the fact that Enterprises are sponsoring the project "here and there", "forcing" the kernel's developer to optimize it for E.L. applications and the Linux Desktop sucks as a consequence of this world-wide behave.

The ONLY "blocking thing" for ROL about to release low-level API to everybody could be (if any) related to technical issues and costs related to make the ROS6's low-level features be able to work also on other ROS branches, but that's all.

Phantomz on 28/7/09 2:49PM
[ Reply | Permalink | Report ]

Yes I would like to see some things I've paid for with the £600+ of Select subscriptions made available, for the good of the platform to encourage application development.

After all I get sod all else out of that investment now, as you refuse to support the only RISC OS machine I still have; the Iyonix, for no good reason.

druck on 27/8/09 12:39AM
[ Reply | Permalink | Report ]

"Remember that there still hasn't been a non-beta release of the 32 bit RISC OS Adjust", you do appear to be skating over things a little bit here.

What the A9home users are waiting on is AD6 to provide the finished OS. Once this has been provided, any updated OS from ROL will softload over it.

sa110 on 13/5/09 2:36PM
[ Reply | Permalink | Report ]

Skating over things? Being 32 bit-neutral is the central flagship feature of RISC OS 6. And nobody's seen it yet. I don't think that's skating over things to point out.

rjek on 13/5/09 3:26PM
[ Reply | Permalink | Report ]

"Skating over things? Being 32 bit-neutral is the central flagship feature of RISC OS 6. And nobody's seen it yet. I don't think that's skating over things to point out."

Obviously you didn't look very hard at the Dual Head A9Home at Wakefield to see what version of RISC OS it was running.

riscosboss on 22/5/09 3:28PM
[ Reply | Permalink | Report ]

Seen it as a product? It's one thing to demonstrate a proof-of-concept prototype, it's quite another to have it as an actual working selling product. Which RISC OS 6 hasn't managed yet, has it?

And the Dual Head A9Home was running with /parts/ of RISC OS 6. The parts that actually worked. Closer examination of the construction of its ROM image was quite revealing.

rjek on 22/5/09 3:56PM
[ Reply | Permalink | Report ]

The dual head A9home was not running full RISC OS 6. It was still several kernel generations behind RISC OS 6.14.

sa110 on 22/5/09 5:27PM
[ Reply | Permalink | Report ]

"Remember that there still hasn't been a non-beta release of the 32 bit RISC OS Adjust that all those RiscPC owners paid for. As for which is the most advanced, well, that's debatable. Most of the new features in ROL's strand are never used."

If you are trying to make out that Risc PC owners paid for the A9Home to be developed, then that is untrue. The conversion of RISC OS 4 to be 32 bit has been part of our business plan since day 1. Advantage 6 paid their own development costs for the parts of the A9Home development that were exclusive to them. If you are trying to make out that we misled people into subscribing to the Select scheme and didn't deliver anything for them, then that is also wrong. In fact most of the things you say are so wrong, that one could laugh out loud, if it wasn't such a cliche, that you yourself constantly use.

I am intrigued to know what your basis is for saying that "most of the new features in ROLs strand are never used." Have you done a survey of users to see if they use image thumbnailing. Or perhaps they don't use DHCP? How about the ability to change the network configuration without rebooting? You seem eager to use the phrase "citation needed" to justify your arguments against us, but where is the proof for your claims that most of the Select features are never used especially since you aren't a Select subscriber.

riscosboss on 22/5/09 4:00PM
[ Reply | Permalink | Report ]

"If you are trying to make out that Risc PC owners paid for the A9Home to be developed, then that is untrue." I am not. I know who paid for the A9Home to be developed.

I gave a list of features I believed most users would never use, or even understand. Of course, you can name a handful of ones that are used, that's easy. Nobody's piped up and said "I use this obscure, ill-thought-out feature!", and they're hardly ever mentioned on usenet, the usergroups I attend, on Drobe, Iconbar, IRC, etc etc etc. Or do you seriously think people often use Sun Raster import? How many people, do you think, would rather effort spent on features like Unicode font support than import support for an ancient and esoteric image format?

I say "citation needed" because you make outlandish claims and refuse to detail why you believe them to be true. At least I give examples for which nobody tends to argue with. And when you /do/ link to another page, most of them time it says the opposite. Just like tlsa said, although you continue to ignore him.

rjek on 22/5/09 4:21PM
[ Reply | Permalink | Report ]

"Nobody's piped up and said "I use this obscure, ill-thought-out feature!""

That's because IMO, none of the feautures of RISC OS 6 are either obscure or ill-thought-out. Perhaps if you used RISC OS 6 on a regular basis, you would see this.

sa110 on 22/5/09 5:32PM
[ Reply | Permalink | Report ]

"That's because IMO, none of the feautures of RISC OS 6 are either obscure or ill-thought-out."

And how many of the ones I listed do you use, or know of anybody who uses them?

rjek on 22/5/09 5:44PM
[ Reply | Permalink | Report ]

Just because nobody comes along and says "I use the filer thumbnailing all the time", or "The ability to import and export pngs in Paint is fantastic", is not evidence of people not using them.

How often do you here of an Iyonix user saying "Aint the DHCP in RISC OS 5 fantasic, I love the way it causes me a headache when the network is not connected"

The point being, just because someone does not go on and on about a particular feature(s) does not mean they are not in use.

sa110 on 22/5/09 6:41PM
[ Reply | Permalink | Report ]

"Just because nobody comes along and says "I use the filer thumbnailing all the time", or "The ability to import and export pngs in Paint is fantastic", is not evidence of people not using them."

Aren't you just falling into the same trap riscosboss did, and ignored the earlier list of obscure additions rjek gave as examples?

VinceH on 22/5/09 7:48PM
[ Reply | Permalink | Report ]

"Aren't you just falling into the same trap riscosboss did, and ignored the earlier list of obscure additions rjek gave as examples?"

I don't think so. Just because rjek doesn't think people use these features, does not mean that they do not get used. Let us not forget all the various features of Select that have been added that rjek did not mention. Not to mention thsoe users who prefer to keep out of these types of discussions on both webistes such as Drobe and other discussions lists.

Built-in USB - only reason it is built-in is because it is Castles USB implementation. RISC OS 6 does not require this to be built into the ROM images due the drivers being provided by a third party, when they purchase a suitable addon card

Unicode - Agreed

PCI Support - There is no system that RISC OS 6 runs on that required this, therefore why include it?

sa110 on 22/5/09 8:13PM
[ Reply | Permalink | Report ]

> Built-in USB - only reason it is built-in is because it is > Castles USB implementation. RISC OS 6 does not require > this to be built into the ROM images due the drivers > being provided by a third party, when they purchase a > suitable addon card

But I thought the A9 had built-in USB ports?

adh1003 on 22/5/09 9:31PM
[ Reply | Permalink | Report ]

Correct, and has rjek and many others over the past few months have pointed out, the A9home is still awaiting it's final OS (i.e. not running RISC OS 6 yet). Whether this will have USB and HID drivers in ROM remains to be seen. I'd hope it would.

sa110 on 22/5/09 9:45PM
[ Reply | Permalink | Report ]

"PCI Support - There is no system that RISC OS 6 runs on that required this, therefore why include it?"

Because, err, without it, it won't run on systems that require it?

rjek on 23/5/09 5:32PM
[ Reply | Permalink | Report ]

There is no system that runs RISC OS 6 that needs it. If it was required, I'm sure it would be added.

sa110 on 23/5/09 6:39PM
[ Reply | Permalink | Report ]

Writing support for hardware is usually a good first step to getting an OS to work on it. Given how simple PCI is really, adding it would open up a lot more SoCs for RISCOS Ltd. not to port to.

rjek on 23/5/09 8:27PM
[ Reply | Permalink | Report ]

I think ROL will have to do something about the ARM netbooks if they do come out as otherwise it focusing on old hardware or emulation. How much of challenge would be for ROL to get there strand to work on the cortex ARM chips.

kuliand on 11/5/09 5:10PM
[ Reply | Permalink | Report ]

How would new user be able to buy in and use something on old hardware when its not really available to buy new, is it ? RiscPC's are so old and available to buy secondhand. As people have pointed out using it on newer hardware in netbooks are really attractive. Notebooks and laptops are what people want to use ?

Hairy on 11/5/09 5:23PM
[ Reply | Permalink | Report ]

"Notebooks and laptops are what people want to use ?"

I agree entirely. I am a supporter of ROL. However the RO5 stuff is looking very good at the moment to the point of eclipsing RO6. I think whilst ROL are still relevant in todays market, their relevance in tomorrows market is much less assured.

However we also need to look at RO5. How long will the current coders continue spending their free time on this project? Will it go beyond the Beagle Board? Will CTL come back with a Beagle based product?

sa110 on 11/5/09 8:20PM
[ Reply | Permalink | Report ]

And how long will ROL continue to get enough money to pay for work on RISC OS? Precisely the same problem exists there, too. At least when the source is available to all, somebody can pick it up again years later when the fancy takes them, rather than it languishing closed, and lost forever.

rjek on 11/5/09 8:30PM
[ Reply | Permalink | Report ]

You are entirely correct of course.

No matter which version of the OS you use, neither is guaranteed to be updated in the future.

sa110 on 13/5/09 2:45PM
[ Reply | Permalink | Report ]

Yes new blood will need to be found over time and how this will be achieved is yet to be discovered but as the ROOL effort is gaining ground it should be easier to get people interested in helping. One of the positives over working on RISC OS over linux currently is the fact that it is a small community so you really can make a difference where that 1000's of people working on linux so how much of impact can one person really have.

kuliand on 11/5/09 9:14PM
[ Reply | Permalink | Report ]

ROL tried.

What the feck happend to RON (Risc-OS on Netbook). If a few guys can get a desktop booting on a beagle board this quickly how come ROL couldn't get RON out the door?

There were a hell of a lot more people left in the market back then and so many people were saying I want risc os on a laptop.

I fear ROL won't be able to get Risc OS 6 running on a cortex netbook any time soon (thought I guess the HAL will offer a easier route than they had back when RON was going to happen).

Hope I am wrong. Agree with the things said in this article.

John

mrmac on 13/5/09 12:53PM
[ Reply | Permalink | Report ]

"What the feck happend to RON (Risc-OS on Netbook)."

I believe the problem with the netbook project was that PSION didn't know a clue about their product, because they didn't actually make it. It was done by another company for PSION.

sa110 on 13/5/09 2:31PM
[ Reply | Permalink | Report ]

Psion made the Netbook. While they had outsourced the design of the "Psion LX" netbook, this project was canned when Psion decided to entirely leave the PDA market and concentrate on the Teklogix business group.

rjek on 13/5/09 3:25PM
[ Reply | Permalink | Report ]

Your knowledge on all these things is impressive!

sa110 on 13/5/09 3:29PM
[ Reply | Permalink | Report ]

This is what happens when you're actually directly involved.

rjek on 22/5/09 1:11AM
[ Reply | Permalink | Report ]

I know you are involved with NetSurf development, but are you also directly involved with RISC OS 5 or 6 development?

sa110 on 22/5/09 9:48PM
[ Reply | Permalink | Report ]

In the OS? Only in the same way a book reviewer is involved in writing it. The hardware, however, is a different matter completely.

rjek on 23/5/09 12:19AM
[ Reply | Permalink | Report ]

That is correct. When we first met with Psion they were very enthusiastic. However we hit the same problem as the Linux developers encountered when they were trying to do a Linux port, in that the NetBook had a custome designed interface chip which had been designed for Psion by a third party. We never managed to get technical specs for that chip, as the company concerned wouldn't release the specs. Psion then announced that they were moving onto the NetBook Pro which was to use a new chip set and was to use Win CE instead of EPOC. By then the investor who had funded initial development of Ron pulled out. Psion themselves hit problems and changes of ownership meant that Psion Teklogix as it became known, no longer wanted to help support a different OS. They did offer to sell us all the remaining stock of the NetBook, but as we didn't have the resources to try and reverse engineer the chip, the project was dropped.

The point is that RISCOS Ltd itself was not set up to try and replace Acorn. Its business plan was to help the existing Software and Hardware Developers continue their businesses by offering a continued availability of RISC OS development, not to offer products that would compete with them.

RISCOS Ltd has continued to do that over the past 10 years, and has put far more into the development of new features and bug fixes for RISC OS than ROOL has. The splitting up of Kernel into separate modules is a case in point as it makes the support of new hardware far easier with RISC OS 6, than using RISC OS 5.

The point is that it is not up to RISCOS Ltd to release new computers using RISC OS 6. It is up to our licensees to do that. We already have in place several RISC OS Licencee Agreements, (including one with Castle) and I refer you to the FAQ and Press Release put out by Castle in July 2003, which confirmed that RISCOS Ltd was the company to contact if anyone wanted to licence RISC OS for desktop computers. Our rates are very reasonable, though of course subject to individual negotiation, based on quantity volumes.

riscosboss on 13/5/09 3:30PM
[ Reply | Permalink | Report ]

What did the interface chip in the original netbook do? Incidentally, Linux runs fine on that hardware, and has done for some time.

"The point is that RISCOS Ltd itself was not set up to try and replace Acorn. Its business plan was to help the existing Software and Hardware Developers continue their businesses by offering a continued availability of RISC OS development, not to offer products that would compete with them." Mentioning no magazines, thumbnailing filers, internet suites etc.

"RISCOS Ltd has continued to do that over the past 10 years, and has put far more into the development of new features and bug fixes for RISC OS than ROOL has." No doubt. But then ROOL's not been around long, and the combination of Pace and CTL have put a lot of effort into bug fixes. Spending effort on "new features" isn't useful if the feature's aren't.

"The splitting up of Kernel into separate modules is a case in point as it makes the support of new hardware far easier with RISC OS 6, than using RISC OS 5." Given amateurs can port RISC OS 5 to completely new hardware in 6 months of spare time, and nobody has yet managed to release a RISC OS 6/Adjust32 system for anything in years, I strongly doubt this. In fact, the way the OS in the A9 is split up, it actually breaks more than it helps.

"The point is that it is not up to RISCOS Ltd to release new computers using RISC OS 6." So, when can we see a 32 bit RISC OS 6 for the RiscPC? This comes with huge benefits at the cost of not running antique software; and nobody would be forced to upgrade, anyway.

rjek on 13/5/09 3:46PM
[ Reply | Permalink | Report ]

"So, when can we see a 32 bit RISC OS 6 for the RiscPC? This comes with huge benefits at the cost of not running antique software; and nobody would be forced to upgrade, anyway."

Personally, I cannot see any benefit of this to the Risc PC user. The downfall of a 32bit RISC OS 6 on a Risc PC is that all the 26bit software that people still use will no longer work, other than running slower under Aemulor.

I'm sure if lots of people get in touch with ROL and and say they want to run a 32bit version of RISC OS 6 on their Risc PC, then they would make it available.

sa110 on 13/5/09 5:38PM
[ Reply | Permalink | Report ]

As I said, nobody who likes their antique software would be forced; but you seemed to ignore that.

As for the advantages; performance, memory efficiency, and single code base. All big and important improvements.

rjek on 13/5/09 8:36PM
[ Reply | Permalink | Report ]

Good to hear a response from ROL. A question, if I may Paul;

Do you consider it likely that RO6 will find itself onto new hardware in the near future? I understand it isn't your responsibility to make the hardware, but will you be actively be looking for partners who will do this?

Presumably if this isn't the case then you see the company as a being primarily around to provide support for older machines; is this correct?

Regards.

Monty on 14/5/09 12:02AM
[ Reply | Permalink | Report ]

"The point is that RISCOS Ltd itself was not set up to try and replace Acorn. Its business plan was to help the existing Software and Hardware Developers continue their businesses by offering a continued availability of RISC OS development, not to offer products that would compete with them. "

To a large extent, isn't this a mute point these days? There is only one hardware manufacture, and as Rob has pointed out, they have been unable to release RISC OS 6 for their product during the last couple of years. Their product whilst satisfactory for many users, is still not exactly cutting edge hardware and not quick enough for other users.

We have more and more ARM based products being or about to be released that would be capable of running RISC OS, but what do we see from ROL, nothing, other than hiding behind the orginal goals of the company.

I want ROL to survive, I think RISC OS 6 is a good product. However unless you change and bring RISC OS 6 to newer hardware, all the features in the world are not going to make people continue running it.

The challenge for ROL, is to show us what you can do. If RO6 is easier to port to newer hardware than RO5, then show us. Enthusiasts have already got RO5 running on newer hardware, from 6 months of dedicated work in their spare time. Stop improving RO6 for a little while (yes I am a Select subscriber and do approve of my subs being used for this), and show us RO6 running on a modern none emulated platform. Not only that, provide a CD with an installer, so that I can buy the OS from you, and the device from someone else.

RISC OS Ltd's website informs us you are " .. soon to celebrate the 10th Anniversary of the launch of RISC OS 4". Why not celebrate it in style with the launch of RISC OS 6 for one of many ARM powered devices that everyone is talking about.

Think out of the box.

sa110 on 15/5/09 3:19PM
[ Reply | Permalink | Report ]

"Think out of the box." There's nothing out of the box about what you've written; it's all good common sense that seems to elude ROL.

rjek on 15/5/09 3:55PM
[ Reply | Permalink | Report ]

"To a large extent, isn't this a mute point these days?"

Presumably you mean a moot point rather than a silent one, unless you were deliberately putting a fresh spin on a hackneyed phrase.

moot –adjective 1. open to discussion or debate; debatable; doubtful: a moot point. 2. of little or no practical value or meaning; purely academic. 3. Chiefly Law. not actual; theoretical; hypothetical.

mute –adjective 1. silent; refraining from speech or utterance. 2. not emitting or having sound of any kind. 3. incapable of speech; dumb. 4. (of letters) silent; not pronounced. 5. Law. (of a person who has been arraigned) making no plea or giving an irrelevant response when arraigned, or refusing to stand trial (used chiefly in the phrase to stand mute). 6. Fox Hunting. (of a hound) hunting a line without giving tongue or cry.

thesnark on 20/5/09 6:03PM
[ Reply | Permalink | Report ]

You may say it is not up to ROL to make new machines available. But if only STD are working on hardware. And there are plenty of suitable netbook/beagle boards etc. machines available or on way. Why would ROL not want to make an OS that people can load onto new, fast and up to date portable hardware.

Saying it is up to a licensee seems very short sited because if no licensee decides to offer new hardware and you'll have no customers to sell new copies of the OS to.

Unless ROL want to die they need to take an active role in either making the OS work on new (and cheap) hardware. Or they need to open up all details on the HAL and OS building so ports can be done. After all I am sure the A9 market is not huge and the RPC one must be shrinking.

John

mrmac on 14/5/09 11:36AM
[ Reply | Permalink | Report ]

"After all I am sure the A9 market is not huge and the RPC one must be shrinking."

But which is reaching their end of life first; The user or the Risc PC?

sa110 on 15/5/09 4:28PM
[ Reply | Permalink | Report ]

I thought ROL did not use a HAL for their versions of RISC OS, like Castle did for RO5/Iyonix.

Don't they still have to make their OS work directly with the hardware, like before the Iyonix came out.

Isn't this why they do not have an OS version that will work on current and all the older computer models?

ajb on 14/5/09 6:23PM
[ Reply | Permalink | Report ]

They don't have a module called "HAL", if that's what you mean. They have a HAL, by virtual of splitting all the hardware access out of the kernel. But it doesn't appear to be as useful, given the number of platforms on which ROL's OS runs on. (ie: two, RiscPC-like and A9Home, verses the at least three of the CTL one, forgetting any boards Pace developed).

rjek on 15/5/09 12:45AM
[ Reply | Permalink | Report ]

You are forgetting that ROL itself is not about putting their branch onto other hardware(perhaps this is something else that needs to change. They don't have to be hardware company to supply an os on cd that works on a third party board). That only happens when an appropriate third party company comes along and wishes to use the OS. Therfore direct comparision, unless you have seen the source for both is perhaps miss-leading.

sa110 on 15/5/09 7:49AM
[ Reply | Permalink | Report ]

You're suggesting that RISCOS Ltd.'s business model is to not put their product into new markets? No wonder we're in trouble. They need to *sell* their product. If they're unwilling to do the work themselves in order to sell more copies of their product, they should be trying to convince hardware manufacturers to use them. That second option will be tricky. There's no information at all about licensing costs and terms on their website.

If they're actually interested in selling more copies, perhaps they should provide a copy of the sources to the RPCemu team, in order to improve the emulator and sell more copies of the "Virtually Free".

However; think about the A9 Home and the Iyonix. The A9 Home still doesn't have a finished OS, and many of the reasons given by A6 for this is the time, difficulty and moving goalposts involved. Hardly an indicator of a good HAL. Then you have the Iyonix that ROL promised an OS for. Again, not a sausage. I think that points very strongly to it not being very good, no?

rjek on 15/5/09 9:35AM
[ Reply | Permalink | Report ]

In Reply "Then you have the Iyonix that ROL promised an OS for. Again, not a sausage. I think that points very strongly to it not being very good, no?"

Or perhaps RISC OS land politics to be more precise.

There are nice touches in RO6 and the Select branch like the animated menu's when you open them up close to the edge of the screen, better support of image filing etc but as you say a bit lost if no one uses the added features and even more of a loss if there is no new hardware to work on but I guess ROL know this hence the emulation route they have taken.

I think ROL have done things well in a market that doesn't have great funds but I think as SA110 says they need to start thinking out of the box a littel more.

bluenose on 15/5/09 6:26PM
[ Reply | Permalink | Report ]

No, I didn't say that RISCOS Ltd's business model is not to put our products into new markets. That is your deliberate mis-interpretaion! Our business is definately to find new manufacturers to use RISC OS, but our markets are restricted by our original Head Licence with Element 14. That Licence (as was clearly stated in the original Press Release from RISCOS Ltd/Acorn - copy at [link]) was exclusive to us and was primarily targetted at the desktop market. Pace had their own markets, e.g Set Top Boxes, and Network Computers. Anything else was fair game.

We don't have the resources to be in the business of trying to predict which processors any new manufacturer will use for future products. As has been pointed out by many people, there are significant differences between different ARM processors, which means a lot of work to support different cores and SoC designs. So if, for example, we had produced RISC OS 4 for the XScale as a test bed, who is to say that anyone would have actually have decided to use an XScale based design? In the event Castle didn't have the expense of developing RISC OS 5 for the Iyonix. They just nicked it from Pace. So we could have spent a lot of time and money making RISC OS 4 work on the XScale and then not have had any customers for it. In comparison Advantage Six came along and wanted to use the ARM9 and so we were happy to work with them on a fixed design.

There is no information publicly available for licensing costs, because it is not a matter for public discussion. It is a matter for individual discussion, based on the customers requirements.

As has been pointed out many times, RISC OS Six does not use a HAL, in the way that the Iyonix does. So your implied comment that RISC OS Six doesn't have a good HAL is meaningless. RISC OS is actually far better abstracted than RISC OS 5.

I would finally point out that RISCOS Ltd never promised an OS for the Iyonix. The availability of Select for Iyonix was dependent on a number of factors. Firstly the availability of the requisite technical information from Castle. (Which was never forthcoming). Secondly co-operation with Tematic who were supposed to have been working on the required RISC OS 5 kernel changes to support the desktop features of Select (this never happened because Tematic was closed down and all the staff laid off when Castle failed to get the sales of RISC OS 5 that they had expected into huge new markets.) So we can claim success in that we signed up 4 manufacturers to use RISC OS 4. Perhaps you could list the manufacturers who have produced products using RISC OS 5? Finally on a commercial level only around 110 people out of a potential total Iyonix market of around 1,000 users actually committed to wanting Select on the Iyonix. That was not sufficient to justify the investment in trying to do all the work ourselves. After all who would have benefitted most? I would suggest that Castle would have expected to make far more profit out of increased sales of the Iyonix, if the far superior Select had been available for it, than we would have made from selling a couple of hundred Select subscriptions. This ignores the fact that many Iyonix users were already Select subscribers, and hence were not about to pay again for Select for Iyonix.

In comparison we have around 5,000 users running RISC OS 4 or 6 on VRPC, and around a similar number of RISC OS 4 users on Risc PC and A7000. So it is clear where our efforts were best devoted. Events have subsequently proved that to have been the right decision, as there will never be any more Iyonix sales. In comparison the ARM NetBooks offer the best option for the future, and there is no reason why we can't produce RISC OS Six for one or more of them, if the RISCOS Ltd shareholders decide that we should go down that route. I should of course point out that no-one other than RISCOS Ltd can licence RISC OS for use on Desktop Computers. (see the Castle FAQ at [link]) That was a key part of our Head Licence Agreement. We have nothing against RISC OS Open offering the Iyonix sources for free and developing new versions of RISC OS 5 for the Iyonix, but if anyone wants to produce a commercial version of RISC OS for a desktop computer then they must come to us for a licence. The costs will be very reasonable, and will not have any commercial impact on the price of a product. You will notice also from the above Castle Press Release that they admitted that RISC OS 5 included most of the RISC OS 4 enhancements, when RISC OS 5 was released. Our Head Licence gave us the copyright to all derivative works created from the sources we received from Element 14, for use in our market after a 4 year period. Q.E.D RISC OS 5 is a derivative of RISC OS 4, and is therefore our copyright, with a free perpetual licence back to the Head Licence holder for use solely in their non-competing products outside of our target market. We want RISC OS to continue to flourish, but it has got to be done in a properly controlled way. If everyone expects RISC OS to be given away for free then who is going to make any money to support new development? How will RComp or CJE or Martin Wuerthner make any money to develop new products, if everyone expects everything RISC OS related to be given away for free? We all need to support the market, instead of trying to put it down. We put a tagline on our new web site to say that it was best viewed by Netsurf, as a clear sign of support for people who may not have heard of Netsurf to give it a try, but you just complained. You don't include any links to RISCOS Ltd on the Netsurf pages, as if to try and make out that we don't exist. How about practising a bit of all round promotion for RISC OS, rather than just links to Drobe, ROOL and riscos.info on the Netsurf browser default page?

riscosboss on 21/5/09 11:50PM
[ Reply | Permalink | Report ]

Maybe people haven't been linking to the RISCOS Ltd. website because up until very recently the front page was simply horrible to look at, the site is still a terrible mess and it has been poorly updated?

On a positive note though, it's good to see it's being redesigned and the new design actually looks quite good, (except for the blue cog, blue on black or gray doesn't work too well).

Gulli on 22/5/09 12:19AM
[ Reply | Permalink | Report ]

"No, I didn't say that RISCOS Ltd's business model is not to put our products into new markets. That is your deliberate mis-interpretaion!" I think Rjek was refering to what Paul Stewart said, not you.

"We put a tagline on our new web site to say that it was best viewed by Netsurf, as a clear sign of support for people who may not have heard of Netsurf to give it a try, but you just complained." I haven't had the opportunity to test it on Netsurf, but perhaps given the suposedly imperfect layout it could have been worded differently. I agree that riscos.com should probably be listed on the netsurf website though.

I suspect that many readers broadly appreciate the points you raised here. I'm not sure I buy the 'free everything' cure all approach either, for example.

Whilst back on the (seemingly never ending) subject of ROL and CTL licences, does 'desktop' extend to cover netbooks and other laptop-a-likes? Might netbooks fall outside of your licence?

Monty on 22/5/09 12:22AM
[ Reply | Permalink | Report ]

"RISC OS is actually far better abstracted than RISC OS 5." Citation needed. RISC OS 6 has ripped out loads of functionality from the kernel into modules, but most actual hardware access has always been in modules, and what wasn't is mostly done through the HAL in RISC OS 5. What remains in the kernel is largely conditionally assembled; which given the source is available is not a problem. Ripping stuff out into modules does not make a portable OS.

"Perhaps you could list the manufacturers who have produced products using RISC OS 5?" How about the countless hardware platforms that Pace used it on? RISC OS 5 already runs on more hardware platforms /now/ than RISC OS 6 does.

"In comparison the ARM NetBooks offer the best option for the future, and there is no reason why we can't produce RISC OS Six for one or more of them" Except for your previous arguement about how you don't want to port to something without guarantee of supply.

"I should of course point out that no-one other than RISCOS Ltd can licence RISC OS for use on Desktop Computers. (see the Castle FAQ at [link])" Again, citation needed. Additionally, which element of the FAQ are you refering to?

"The costs will be very reasonable, and will not have any commercial impact on the price of a product." If that is the case, I'm sure you'll be happy to publish this information.

"You will notice also from the above Castle Press Release that they admitted that RISC OS 5 included most of the RISC OS 4 enhancements, when RISC OS 5 was released." And how many of them were ROL's doing? I note that the most obvious of them, the 3D window hack, is missing.

"Our Head Licence gave us the copyright to all derivative works created from the sources we received from Element 14, for use in our market after a 4 year period." The legal standing of this statement is simply astonishingly wobbly, as well as vague as can be. Firstly, it refers to the sources you received, not the sources Castle received. Past that, you simply can't assign copyright automatically like that.

"If everyone expects RISC OS to be given away for free then who is going to make any money to support new development? How will RComp or CJE or Martin Wuerthner make any money to develop new products, if everyone expects everything RISC OS related to be given away for free?" I'm sure the free software movement has passed you by; even OSes that aren't themselves open and free have thriving application businesses, both open or otherwise.

"We put a tagline on our new web site to say that it was best viewed by Netsurf, as a clear sign of support for people who may not have heard of Netsurf to give it a try, but you just complained." That's because it wasn't best viewed in NetSurf; it was dreadful HTML exported by a Windows application that made assumptions that are mostly untrue under RISC OS.

"You don't include any links to RISCOS Ltd on the Netsurf pages, as if to try and make out that we don't exist. How about practising a bit of all round promotion for RISC OS, rather than just links to Drobe, ROOL and riscos.info on the Netsurf browser default page?" Because your website has no content of interest, and most users wouldn't feel the need to visit it. We provide links to several RISC OS-related enterprises; mostly those who directly support and sponsor us. Perhaps you'd like to make a contribution.

In all, this post's the funniest thing I've read all week; circular arguments, vague assertions, and outright self-contradiction. Well done, Paul. I'm sure even fewer people now take you seriously.

rjek on 22/5/09 12:47AM
[ Reply | Permalink | Report ]

"Perhaps you could list the manufacturers who have produced products using RISC OS 5?" How about the countless hardware platforms that Pace used it on? RISC OS 5 already runs on more hardware platforms /now/ than RISC OS 6 does.

I think your response needs some serious citations. Perhaps you could answer my original question. i.e how about listing the actual products that Pace produced that used RISC OS 5, and more interestingly those that come up with any form of banner or copyright attribution to Castle Technology and state that the base OS of the product is RISC OS 5 copyright Castle Technology 2004 (or later).

riscosboss on 22/5/09 12:57PM
[ Reply | Permalink | Report ]

We're talking about ease of portability. And ROOL's version already runs on more platforms that ROL's efforts. Here is where the story ends. I've witnessed Pace's RISC OS running on several different pieces of hardware over the years.

Nothing you've cited proves your ownership argument at all, as tlsa touched on a moment ago.

rjek on 22/5/09 1:32PM
[ Reply | Permalink | Report ]

"We're talking about ease of portability. And ROOL's version already runs on more platforms that ROL's efforts. Here is where the story ends. I've witnessed Pace's RISC OS running on several different pieces of hardware over the years."

Ah Yes, you've seen in running on Pace's RISC OS. e.g DSL4000. But can anyone actually list a non-Castle product running "Castle's" RISC OS 5! At present RISC OS 5 runs on Iyonix, and the first stages of the Beagleboard, plus possibly the A7000, and possibly RPCEmu I make that three/four platforms.

RISC OS 4/6 has currently been shown running on ARM 6, ARM710, StrongARM, ARM9, ARM7500, VRPC and RPCEMu. It also works directly with Viewfinder graphics cards (something that Druck once claimed we would never be able to do.) and also the new Vpod graphics cards. That makes quite a few choices of platform, oh and I haven't included the Omega with its own graphics chip. How long will it take for RISC OS 5 to run on a Kinetic Risc PC, with a Vpod attached? That will show how good the RISC OS 5 HAL is!

Nothing you've cited proves your ownership argument at all, as tlsa touched on a moment ago.

riscosboss on 22/5/09 3:26PM
[ Reply | Permalink | Report ]

Firstly, we can assume that the RiscPC-class machines are essentially one and the same, beyond a few minor tweaks for the CPU. That means ARM 6, ARM710, StrongARM, ARM7500, VRPC, and RPCemu are all the same thing. There has been no version of RISC OS 6 released for ARM9 machines, and only an unstable beta for RISC OS 4. So that's one machine, perhaps one and a half if you include the A9. RISC OS 5, however, runs on the Iyonix (with graphics card support, like you claim for the VPod), RiscPC-class hardware (half a point), the BeagleBoard (half a point), and all of the different hardware Pace used it on (we'll call this zero points, but pointing to its working). It'd also most likely work under VRPC too if it hadn't been intentionally nobbled to only accept ROL ROM images. That's two verses ROL's 1.5, and that's being mean to the RO5 side.

RO5 *already* supports different video cards. It *already* runs on more architectures. And you've completely ignored tlsa's point that nothing you link to proves, or even agrees with anything you're saying at all.

You seem to be fundamentally confused between the CPU and the system, as well as what a HAL is. The RiscPC is one system, which a choice of (very similar) CPUs. And I don't doubt that a graphics driver for the VPod couldn't be written for RO5 pretty trivially; the SM501's quite easy to drive simplistically.

C-. Must try harder to understand the issues.

rjek on 22/5/09 3:54PM
[ Reply | Permalink | Report ]

Rob, I gave a simple question which you have still not answered. Please list the products (other than the Iyonix) which Pace (or anyone else) have produced which use a Castle developed RISC OS 5. Of course citations are needed if we are playing that game. You keep on claiming that Pace used it on lots of different hardware, but lets have a list and some evidence that it is RISC OS 5, and not a Pace branch.

riscosboss on 23/5/09 11:12AM
[ Reply | Permalink | Report ]

That question is not relevant to the discussion. The discussion was about how hard it is to port your OS; and there is good evidence to suggest that RO5 is easier, simply because its historically managed to run on more platforms, even excluding the platforms that Pace ran it on before selling their rights to Castle.

This conversation is now serving no purpose. Some might say it is because all it appears you're trying to do is protect your salary, and not further the development of our favourite hobbyist OS.

rjek on 23/5/09 12:28PM
[ Reply | Permalink | Report ]

I assume you're not familiar with the DBxxx series IPTV set-top boxes, with features in the later products like standard definition H.264 decode in hardware. There had multiple ARM cores which were used where possible, e.g. a core was used to independently run a specially modified internet stack. These are of course obsolete products today, since IPTV moves quickly, but bits of information can be found online if you dig around. The DB210 and 220 came first (note the dates in the URLs - this isn't the earliest possible reference on Pace's site, but it's the first I tried after mid-2004, when Pace doesn't list these products at all):

[link] [link] [link]

The IP215 is called an IP215 rather than DB215 because Pace kept changing their minds about naming things :-)

- you'll notice from their earlier products page that they call them "digital broadband media" devices, but later, explicitly refer to them as IPTV boxes and rename the whole RISC OS based range:

[link] [link]

The earlier dated-based link shows the beginnings of the "IP" series naming. As a side-note, the IP4xx and (not listed) IP5xx series were somewhat ill-fated Linux-based designs using some rather unusual hardware. This was the last product I (or anyone else, I guess!) worked on at Pace Cambridge, with a few people kept on for a short while to support the products from a tiny office in Cambourne (IIRC).

Despite the similar names, the older MPEG 2 series differed quite a lot from the later H.264 boxes. Although the ARM architecture therein was compatible with 32-bit RISC OS, the rest of the chipset - for audio, video, infra-red and so-on - bore little resemblance to the likes of the ARM 7500FE-based DSL 4000.

ROOL brought along a couple of these STBs to Wakefield 2007, ROOL's first ever show:

[link]

Now, you will note that these products were appearing on the Pace site during 2005. I worked for Castle at the time and helped write some of the software inside them. I even went to IBC and demonstrated the products. Since Castle bought RISC OS from Pace in 2003 - according to press releases on both Castle's web site and Pace's web site - surely you can accept that the OS in these 2005 products was, indeed, "Castle" RISC OS?

[link] [link] [link]

There is surely no such thing as a "Pace branch" versus "Castle developed RISC OS 5". There is just Castle's RISC OS which was purchased from Pace. I don't see how there can possibly be any confusion or ambiguity here.

adh1003 on 23/5/09 3:51PM
[ Reply | Permalink | Report ]

Quote from John Kortink: "If so, be advised that only Rage 128 Viewfinders appear to be supported under RISC OS 6. Contact RISC OS Ltd, they are responsible for that."

If this is true, when will Radeon Viewfinders be supported under RISC OS 6?

egel on 22/5/09 4:03PM
[ Reply | Permalink | Report ]

"ARM 6, ARM710, StrongARM, ARM9, ARM7500, VRPC and RPCEmu"

I really hope that you don't consider this a demonstration of portability. And even then, if you want to draw up a list of CPU core compatibility, I think you'll find the Pace/Castle fork wins. From the list in the sources:

ARM 600, ARM 610, ARM 700, ARM 710, ARM 710a, SA-110, ARM 7500, ARM 7500FE, SA-1100, SA-1110, ARM 720T, ARM 920T, ARM 922T, 80200, 80321, Cortex-A8

Or in terms of CPU architecture, your supported list is

ARMv3, ARMv4, ARMv4T

whereas Pace/Castle support

ARMv3, ARMv4, ARMv4T, ARMv5TE, ARMv6, ARMv7

But as I'm sure you appreciate, this is only a small part of the story. Granted, you have support for the ViewFinder and the Omega graphics chip, but at least as much effort goes into driving the rest of the I/O system, of which you only appear to support two varieties, IOMD and S3C24xxx (and even then, apparently, tragically, not via a unified source tree).

By comparison, the Pace/Castle fork has support for at least 7 (that I know of) radically different I/O architectures - not counting support for complex external silicon like Chrontel, DENC, MPEG or audio CODEC chips. And the HAL was designed as a separate binary component to enable OEMs to write support for their own I/O chipsets without having to pass the source back to Pace, so there may be others that I've never seen.

bavison on 23/5/09 6:36PM
[ Reply | Permalink | Report ]

Quoting riscosboss "I should of course point out that no-one other than RISCOS Ltd can licence RISC OS for use on Desktop Computers. (see the Castle FAQ at [link])"

"Again, citation needed. Additionally, which element of the FAQ are you refering to?"

Easy. The bit in the FAQ that doesn't quite say what he's claiming. From the page he linked to at [link]

"Q. I want a license to use RISC OS - must I apply to Castle to get one? A. Not necessarily. A number of companies have the ability to grant 3rd party binary licenses for RISC OS, for example RISCOS Ltd can licence RISC OS 4 for desktop computers. In addition they could also apply to Castle for an extension to include new hardware - there are no plans at all to withhold permission to any genuine volume applicant."

Interestingly, I've just looked at an old letter I received from Acorn back in 1992 and by reading and interpreting it in a similar way that other documents appear to have been read and interpreted, I conclude that Acorn actually handed all rights to RISC OS to me, commencing 15 years after the date of that letter. That would be 12th May 2007.

Honestly. The letter includes my name and address, the words "RISC OS", "licence", "transfer", "fifteen" and "years" in that order, as well as many others inbetween them. But I can't show it to anyone.

I therefore hereby assign all rights to ROOL.

Problem solved.
VinceH on 22/5/09 7:39PM
[ Reply | Permalink | Report ]

Oops. Sorry about the italics.

VinceH on 22/5/09 7:40PM
[ Reply | Permalink | Report ]

I suggest you take Castle and ROL to court right away.

Monty on 22/5/09 9:20PM
[ Reply | Permalink | Report ]

"RISC OS is actually far better abstracted than RISC OS 5"

The problem you have here Paul, is that many people for whatever reason don't think it is. One the big reasons for this is that more than a year after the RISC OS 6 release, the only manufacture of a RISC OS based computer has still been unable to release a version of RISC OS 6 for their product. Assuming RISC OS 6 was easier than RO5 to port on to new hardware, surely a version of 6 should have appeared by now?

Perhaps you could direct ROL's resources, for a couple of months, to help get RO6 on the A9home.

"In comparison the ARM NetBooks offer the best option for the future, and there is no reason why we can't produce RISC OS Six for one or more of them, if the RISCOS Ltd shareholders decide that we should go down that route."

Perhaps the RISCOS Ltd shareholders need a reality check! I know you have tried to get them to allow RISCOS Ltd to do other things in the past. Perhaps you could try again. I'm sure if you need help in convincing them, there will be a fair few users who would glady help you make the case to the shareholders.

sa110 on 22/5/09 8:15AM
[ Reply | Permalink | Report ]

"We put a tagline on our new web site to say that it was best viewed by Netsurf, as a clear sign of support for people who may not have heard of Netsurf to give it a try, but you just complained"

I'm sure where I involved in the NetSurf project, I'd be complaining too. Whilst it is good to see the redesign of the RISCOS Ltd website in progress. The site is "not best viewed in NetSurf". Perhaps you should download the latest version of NetSurf and see for yourself the reason why Rob has "complained"

sa110 on 22/5/09 8:21AM
[ Reply | Permalink | Report ]

I consider it to be utterly pathetic that a company can claim a "success" from sitting on it's hands for 6 years ignoring over 1000 potential customers (20% of the claimed total customer base), until the machine they are using finally goes out of production.

BTW the word "nicked" is certainly actionable.

druck on 22/5/09 10:14AM
[ Reply | Permalink | Report ]

"That Licence (as was clearly stated in the original Press Release from RISCOS Ltd/Acorn - copy at [link] ) was exclusive to us and was primarily targetted at the desktop market."

Actually, the press release just says that the licence granted RISC OS Ltd "an exclusive license to complete the development of RISC OS 4". Then it goes on to say ROL intend to flog it for desktop machines, although it says nothing about which companies are licenced to sell in which markets. I don't doubt that there is more to the licence than was stated in the press release, but still it says nothing about anything except that ROL could exclusively complete RISC OS 4.

Out of interest, is that why when ROL leapfrogged RISC OS 5, it was called "RISC OS Six" instead of "RISC OS 6"?

"In the event Castle didn't have the expense of developing RISC OS 5 for the Iyonix. They just nicked it from Pace."

There were press releases on both the Pace site and the Castle site that Castle had licensed RISC OS and then ultimately bought it, iirc. Why do you claim Castle stole it from Pace?

"I would finally point out that RISCOS Ltd never promised an OS for the Iyonix."

Actually, we (my Dad and I) were convinced by yourself at two successive Wakefield shows to part with the Select subscription fee for a product we wouldn't use on the basis of that promise.

"I would suggest that Castle would have expected to make far more profit out of increased sales of the Iyonix, if the far superior Select had been available for it, than we would have made from selling a couple of hundred Select subscriptions."

Of course this overlooks the customers who had already been convinced to subscribe on the basis that their subscription would pay for the work to be done.

As for which version is superior, I guess it depends on which features users find useful. ROLs version has more in the way of decorative trinkets and baubles, as well as some useful stuff like text selection and cut&paste in writable icons. Castle's has the Unicode font manager and a USB stack that seems to have comparatively good device support. Personally, I just think it's a shame that there isn't just one version, with all the good stuff in.

"If everyone expects RISC OS to be given away for free then who is going to make any money to support new development? How will RComp or CJE or Martin Wuerthner make any money to develop new products, if everyone expects everything RISC OS related to be given away for free?"

By offering great products? I help develop NetSurf, which is available for free and I am excited by the prospect of a freely available ROMs from RISC OS open. However, I am always glad to buy an ArtWorks upgrade because I think it's an excellent program which I find very useful and a joy to use.

"We put a tagline on our new web site to say that it was best viewed by Netsurf, as a clear sign of support for people who may not have heard of Netsurf to give it a try"

Thanks for the link.

"but you just complained."

I haven't read the whole thread closely, so I'm not sure what the complaint was. Anyway, I can think of two things worth mentioning:

1. There's a typo in the tagline -- it's "NetSurf", not "Netsurf".

2. The site is not designed to be laid out by the browser in the normal way. Instead of saying "now write render this paragraph of text, wrapping it to fit the width of the container", it says "plot this line of text at x/y coordinate A and then plot this next line at x/y coordinate B". This means that in order to avoid text appearing to overlap out of it's space, the text must be rendered with a font that has no greater width for a particular pt size than the font that the page was designed with. Also RISC OS uses 90 dpi and other OSes use other values like 96 dpi. So such a site design makes it difficult for NetSurf to support on RISC OS, while we continue to use RISC OS fonts and the RISC OS font manager.

"You don't include any links to RISCOS Ltd on the Netsurf pages, as if to try and make out that we don't exist."

Not really. When I made the welcome page, I just anted to include a few useful links for people to try when they first run NetSurf. It was never the intention to make it a comprehensive links database! The four columns of links are roughly: News, IT, Information and RISC OS. I included RISC OS, because I wanted to give the platform some exposure beyond the RISC OS scene (NetSurf is used on other platforms).

Since I didn't want to have vast lists of links which would just put people off looking at them, and I wanted the page to look quite nice and balanced (in terms of layout), I had to be selective. There are three RISC OS links, one to a purveyor of RISC OS, one to a RISC OS news site and one to a site with information useful to people using RISC OS.

I also had some rules on what stuff I did not think should be linked to:

1. No links to sites where the service involves spending (like Amazon or Play)

2. No links to sites where the service requires a login (like Hotmail or Gmail)

As for the reason I chose to link to RISC OS Open's site rather than RISC OS Ltds it was partly reason 1, partly because RISC OS Open's site looked better, was easier to navigate and had a nice straightforward intro to RISC OS on the homepage, and finally because RISC OS Open's purpose and way of working are least different from NetSurf's.

Obviously linking to both is more confusing to people who are new to RISC OS and only serves to highlight the fact the platform is split down the middle and in a bit of a mess!

tlsa on 22/5/09 11:19AM
[ Reply | Permalink | Report ]

Lets not forget the fine job which was done on ViewFinder support in RISC OS 6.

I did have the one variant of card which RO6 supports, and after enjoying 2048x1536x32bpp@70Hz for over 5 years with RO4 and JK's firmware, RO6 decided it could only do 16bpp at an unusable 48Hz.

According to ROL the card can only do 1920x1440x32bpp, so that's all RO6 will allow, and I was just imagining the last 5 years. My response was to bin the entire RISC PC, and versions of the OS which don't run on the far more powerful RISC OS machine that I actually use.

druck on 23/5/09 12:37AM
[ Reply | Permalink | Report ]

sa110>"You are forgetting that ROL itself is not about putting their branch onto other hardware"

That's them limiting themselves. Acorn isn't around to say "Ahem you're not suppose to do that".

I feel (sadly) that as ROL never had (and I suspect never will) have any in house hardware development skillset and that therefore they won't be in a position to port to new hardware anytime soon.

That leaves *emulation*, which is probably the "path of least resistance" that ROL would follow - and continued support for "fossilware" (i.e., RISC PC)

The contrast with RO5 is an interesting one - you could argue that given that Pace (and then later CTL) because they developed hardware better understood the needs of the OS/Hardware and ultimately produced code that was *easier* to port to new hardware (the BeagleBoard port being an effective proof of this).

I suspect ROL will cater for a proportion of RISC OS users - but will not make meaningful inroads onto new hardware or onto new platforms.

If they wished to do this they'd need to think radically - perhaps even licensing parts (or all) of RO5 from Castle on a commercial basis and modifying RO6/Select to work with it. Or alternatively opening their OS so that similar developments to what is happening on the ROOL branch would occur... the problem is that that would deprive ROL of some income and would represent a risk.

ROL have had some success - but to expand from where they are would require different thinking and possibly doing what for them in the past would appear unthinkable (mind you they did this once before when they embraced Windows and Emulation - so perhaps they make an equivalent leap again - who knows).

AMS on 16/5/09 10:33AM
[ Reply | Permalink | Report ]

ROL should have a close look at their website...

With NetSurf (current build) it looks ok but does have some issues. With Phoenix it is a sheer mess and Oregano 2 is not much better. Who is supposed to go there?

hzn on 17/5/09 10:40AM
[ Reply | Permalink | Report ]

Looks fine in IE8. Shame about "best viewed with Netsurf" footnote. Really it should read "Don't view with NetSurf and any other RISC OS browser". Firefox or IE only."

sa110 on 17/5/09 11:07AM
[ Reply | Permalink | Report ]

I don't think it should be forgotten that ROL are the only company actually making any money at the moment out of selling RISC OS.

So it is understandable that they are attending to their existing market first, and would want to think very carefully before doing anything radical which would jepordise this. So, just to gather together my thoughts, by existing market I mean, emulation via Virtual Acorn, finishing the OS for the A9 and incremental progress for Risc PC users via the select scheme.

I have been worried that the 'open' emulators are going to take out one of ROL's income streams. I worry because they'd be forced to defend their market which could result in 'friction'.

Porting RISC OS 5 to, say, the Pandora, is a much better proposition as it does not compete directly with existing markets. It would be fantastic if Castle could be brought back into the game, as a supplier of an "out-of-the-box" Pandora, for example. I'd certainly be very interested in buying such a "user friendly" product. To me, that's a sellable proposition, and it's entirely right that CTL take a cut for the "value added". Possibly, ROL get a cut too from licensing - I don't know about that.

The most important aspect of all of this, is for all parties to respect what each other are doing, not tread of each others toes, and focus on growing the use and usefulness of RISC OS in their own areas.

There are a lot of exciting and interesting things going on in the RISC OS world at the moment and we need to keep it all growing and moving forward.

Regards, Martin Hansen.

martin on 19/5/09 1:47PM
[ Reply | Permalink | Report ]

"I worry because they'd be forced to defend their market which could result in 'friction'" So you don't think there has been much 'friction' recently?

"Possibly, ROL get a cut too from licensing - I don't know about that" As ROL will not put their money where their mouth is, I assume we have to operate under the known facts of:

1)Castle own RISC OS 2)ROL operate under a license from Castle 3)ROOL operate under a licence from Castle

Don't see where "ROL get a cut too from licensing" comes into it.

Your suggestion would have ROL treading on Castle's toes.

ajb on 19/5/09 5:34PM
[ Reply | Permalink | Report ]

Depends on who's view of the contracts you subscribe to. Unless either view actually gets sorted out in court, both views are to a large extent a mute point as long as neither side actually makes a challenge in court.

sa110 on 19/5/09 7:45PM
[ Reply | Permalink | Report ]

Ah.. OK. I thought I read somewhere that those known facts were in dispute but I'm probably (and hopefully) wrong.

Thanks for the correction.

I wasn't quite sure what you meant by "putting their money where their mouth is" but if it means they are careful about what they do with their resources it's probably all to the good.

Regards, Martin.

martin on 19/5/09 7:49PM
[ Reply | Permalink | Report ]

"OK. I thought I read somewhere that those known facts were in dispute"

No, those are known facts. What are in dispute are recent statements made by ROL. That's where the money/mouth comes in.

ajb on 20/5/09 12:03AM
[ Reply | Permalink | Report ]

I don't want to get into this again, but in the interests of completeness, you left out the minor detail that the licence in item 2, ("ROL operate under a license from Castle") is actually an exclusive licence for the RO desktop market. Hence there is reasonable dispute about the legal status of other desktop RO implementations.

Check out the following Drobe article for the latest (unless I missed one) public version of this dispute.

[link]

Stoppers on 20/5/09 6:28AM
[ Reply | Permalink | Report ]

No, under the new 2004 licence agreement (which ROL have officially said that they signed it), the often-cited exclusiveness is no longer. Hence ROL now pretending (see Aaron's last few postings) that this new licence agreement is now somehow non-binding.

However, if everything that has been said by both parties (and third parties) is actually true, the whole situation is a real mess, and it is not surprising that no party want to actually have it tested in court.

hubersn on 20/5/09 11:54AM
[ Reply | Permalink | Report ]

hubersn: Couldn't we say claiming, rather than pretending, please? It's less judgemental.

Aaron also claimed that the 2004 agreement was "abandoned by both parties" (it would have had more impact had he given some more details about what form that agreement took, I would hope for something on paper).

It is indeed a real mess, and I don't know which way a judgement would go, but one thing that there's no doubt about is that the situation is still disputed.

Stoppers on 20/5/09 2:33PM
[ Reply | Permalink | Report ]

What i have heard is that CTL refused to abide by the disupute term, as detailled in the 2004 contract. Whether this is true or not, I cannot say.

sa110 on 20/5/09 5:38PM
[ Reply | Permalink | Report ]

Has the wording of the complete license that ROL operates under ever been made available for public viewing?

If it's as air tight as you believe, ROL would surely have protected itself by using the legal system in the past.

If not it would normally just have to accept the situation, so as to not cause further damage to their own business, or the general business environment. This is what I found most confusing with ROL behaviour.

Ref your link, I would direct you to the "RISC OS Open's response" at the bottom of the article. I think Steve says it all.

ajb on 20/5/09 12:17PM
[ Reply | Permalink | Report ]

"If it's as air tight as you believe, ROL would surely have protected itself by using the legal system in the past."

If it's as air tight as you believe, Castle would surely have protected itself by using the legal system in the past.

As to what has been made public, if you read my contributions to the other thread [link] you will find out what I know, why I know it, and what my opinion is (it may not be what you think, see the penultimate paragraph of [link])

Stoppers on 20/5/09 12:48PM
[ Reply | Permalink | Report ]

"If it's as air tight as you believe, Castle would surely have protected itself by using the legal system in the past".

No, you cannot wriggle out of the position you have taken up, by simply turning the statement without giving any reference. We were discussing the ROL license, and I had not been quoting any unseen Castle licenses.

Castle has, however, used the legal system in the past. Remember back when ROL had a legal restraint put on them.

It is ROL that is making the claims about their license, that you are quoting as though they are airtight. I notice that you do not claim to have seen the license in question.

At present Castle seems to be happy with their current business situation, as they have not been making any claims against anyone, which I have heard about.

I presume that they would only consider initiating legal proceedings about something if they thought they needed to, and that it would be of benefit to their business operation.

I have read your referenced contributions, and it appears to me that you are happy to quote hearsay as though it is gospel.

I initially mentioned 3 straight forward facts, to which you added your comment about the 2nd fact. As you appear to have not seen the license, it you perhaps be better if you didn't quote hearsay as fact.

ajb on 20/5/09 4:37PM
[ Reply | Permalink | Report ]

"Castle has, however, used the legal system in the past. Remember back when ROL had a legal restraint put on them."

You mean the "dial a lawyer" episode to send out legal letters to intimidate your oposition?

From what I've heard on that one, I believe the only thing that achieved was Castle almost being laughed out of the ROL share holders meeting!

sa110 on 20/5/09 6:32PM
[ Reply | Permalink | Report ]

And there was me thinking I remembered that there was a big issue about ROL, and some other companies, having to stop trading to the public for quite a while.

Obviously I was mistaken, "if I blindly accept what you say".

ajb on 20/5/09 6:48PM
[ Reply | Permalink | Report ]

It would be wrong of any company not to check their legal status when Cease and Desist letters are sent out.

However just because company X sends C & D letters to companies A & B, does not necessarily mean company X is in the right either. If A & B take their own legal advise and decide the C & D letter that company X sent out does not apply and continues regardless, it is up to company X to follow through on the C & D letters and take the matter to court.

sa110 on 20/5/09 7:47PM
[ Reply | Permalink | Report ]

"And there was me thinking I remembered that there was a big issue about ROL, and some other companies, having to stop trading to the public for quite a while."

I can't speak for other companies, but RISCOS Ltd received numerous letters from Castle concerning lots of allegations. We NEVER stopped trading, though Castle clearly wished we had. Lots of people on this discussion continue to bring up so called "facts" that are nothing more than hearsay.

The simple truth is that the British legal system is very much split into two strands. Criminal Cases and Civil Cases. In a Criminal Case the law is usually quite clear, and someone who is guilty will usually be convicted. In a Civil Case there is no clear cut distinction between Right and Wrong. It is down to whoever has the deepest pockets to pay for lawyers! i.e you have to be able to pay for your legal costs plus any damages you might become liable for if you were to lose. e.g it makes it very hard for a local computer shop to take on PC World, as the losses that PC World might claim would be huge. However if a new business starts up, then it is a lot simpler and cheaper to take action against them if they are doing something wrong e.g if they start selling something without a licence that is your copyright.

Ten years ago RISCOS Ltd and Element14 entered into an agreement between the two parties. At various stages two further parties have become involved. The original parties are the ones who count in this situation, as the spirit of the agreement will have been determined by them. Castle made certain allegations against RISCOS Ltd based solely on their interpretation of the Agreement. They were not parties to the original Agreement, and hence were on very dodgy legal ground to try and stop our contract, to which they themselves were not an original party.

So the situation at present is that we are still operating under the terms of the original Head Licence, and that gave us very clear rights. The Head Licence Agreement is obviously written in legaleese, but was written following a Memorandum of Understanding between the parties, which gave the conditions in very plain English.

The Head Licence Agreement is however subject to confidentiality and Drobe is not the forum for discussing commercial contracts.

I am however looking to make the Memorandum of Understanding available on the web site, along with supporting documents.

We obvioulsy want to see innovation in RISC OS development, but if anyone starts to encroach on areas where we have exclusivity, or starts using software that is our copyright, then we will have to take action to protect our rights.

There is obviously scope for ROOL and RISCOS Ltd to co-exist and I hope that we can come to an agreement for co-operation, and with safeguards to protect the investment that has been made in RISC OS development over the past ten years.

riscosboss on 2/6/09 1:17PM
[ Reply | Permalink | Report ]

"No, you cannot wriggle out of the position you have taken up..."

The position I am taking is that the licence between ROL and Castle had, at least until 2004, a clause granting exclusive right to RISC OS Desktop products, and that they still claimed that (and more) to be the case, as recently as December of 2008.

Your position seems to be that ROL have no rights over desktop RO products, is that right?

You may have noticed that I've said that I don't know which way a ruling would go, since both sides have been playing silly beggars.

"We were discussing the ROL license, and I had not been quoting any unseen Castle licenses"

There is no difference; the licence is shared by the two companies, with rights and obligations on both sides. (Assuming it has not legally been revoked.)

"Remember back when ROL had a legal restraint put on them."

Vaguely. Was that on the basis of a ruling against them in court, or just a solicitor's letter? Do you see a difference?

"It is ROL that is making the claims about their license, that you are quoting as though they are airtight."

OK, I'll qualify it a little; their licence was an exclusive licence for the RO desktop market; they were recently claiming that hasn't changed.

"I notice that you do not claim to have seen the license in question."

No, I explained in the other thread the documentation I have seen. I assume that Castle made the best argument possible in support of their case. In it they state certain facts about the licence (including a part that ROL held exclusive rights to the market for desktop RO, which was, unsurprisingly, not disputed by ROL) and also arguments as to why Iyonix didn't fall under within that market, which I found it unpersuasive, to say the least.

May I ask what documentary evidence have you been party to, to be so sure that ROLs claims can be so easily dismissed?

"I presume that they would only consider initiating legal proceedings about something if..."

That's why I turned your statement around; why do you think ROL should behave differently?

"I have read your referenced contributions, and it appears to me that you are happy to quote hearsay as though it is gospel."

I'm not convinced by gospel.... :-)

I have not read the licence. On the other hand, I have read Castle's "Briefing to ROL Shareholders" regarding the licence between the two companies. You're welcome to consider that hearsay, if you like.

It's true you mentioned 3 facts, but the way you put it (and the following sentence) gave the impression that Castle had no obligations to ROL, which is something ROL was still disputing a few months ago.

I don't know why you can say you "Don't see where "ROL get a cut too from licensing" comes into it.", with no knowledge of the agreements between the two companies. If you've been told that's the case, that's hearsay, if not, it's just uninformed speculation. IMO, it's more reasonable to expect that ROL should get some recompence in exchange for their presumed relaxation of the licence permitting them exclusive access to the RO desktop market.

I, personally, am sure (based on written documentation) that the licence included obligations on both parties (ROL and Castle, originally E14), at least at the time of the briefing I attended in June 2004 (I notice I mistakenly put 2008 instead of 2004 in the other thread), and ROL, in the person of one of its Directors, was still disputing Castle's assertions in December 2008.

Stoppers on 20/5/09 6:58PM
[ Reply | Permalink | Report ]

Surely thats enough of that. There's plenty of space on c.s.a for tedious, lengthy, hampster-wheel legal arguments without infecting yet another drobe thread with them.

Monty on 20/5/09 7:27PM
[ Reply | Permalink | Report ]

"Surely thats enough of that."

It should be.

I get upset when accused of repeating hearsay by someone who doesn't back up their statements at all.

It's a character flaw. :-(

Stoppers on 20/5/09 7:34PM
[ Reply | Permalink | Report ]

Your position should be that of "on the fence", given neither company has produced any evidence of their point of view; what they have produced is either laughably circular or vague in the extreme. So the solution is to just shut up and ignore the issue until one of them sues the other, or one goes bankrupt.

rjek on 20/5/09 8:08PM
[ Reply | Permalink | Report ]

That is indeed my position, but I will continue to insist that the fence I am sitting on actually exists, when somebody tells me it doesn't.

Shutting up, now! :-)

Stoppers on 20/5/09 8:47PM
[ Reply | Permalink | Report ]

They would be making a bit more money if they replied to my email enquiring about the PRMs manuals.

How fast do they normally respond?

kuliand on 20/5/09 12:33PM
[ Reply | Permalink | Report ]

We responded at 4:39 p.m on Sat 16th May to your original email sent at 17.17 p.m on Friday 15th May. You didn't reply until 13:39 on 20th May. In case you have lost the emails, the Manuals CD can be purchased for £6.50 incl postage. Orders can be placed by email, post or via the website.

riscosboss on 21/5/09 1:36PM
[ Reply | Permalink | Report ]

Ye sorry gmail seems to be playing up with emails taking along time to get through to me.

I'll get and put a cheque in the post.

kuliand on 21/5/09 7:05PM
[ Reply | Permalink | Report ]

edit:

scrap that i have just seen you have account on ebay so i'll buy it through that.

kuliand on 21/5/09 7:36PM
[ Reply | Permalink | Report ]

Besides....

I think it was said ROL has not paid all it's back license fees to castle.

If that is the case then it may be as well if they don't start causing problems.

They may make a current income stream but how long will that exist if we have Iyonix running ROS 5, Beagle Board running ROS 5, Netbooks with ROS 5, emulator with ROS 5. If we have a stable RPC/A7000 build of ROS 5 why would people stick with diffrent versions? (ok so I know some may for certain software or summint) but if we have new fast hardware all on same OS would it not make sense to have older hardware on Same OS as well (would be nice for software developers).

If that happens I am willing to bet unless ROL do somthing before it happens all their income streams will be seriously reduced if not gone. Hence my point they need to either position themselfs to get RISC OS 6 onto portable hardware/new hardware. Or make Select features sit on top of ROS 5 and quick....

John

mrmac on 20/5/09 1:13PM
[ Reply | Permalink | Report ]

Will a 32-bit RISC OS handle 26-bit podule ROMs, though?

If not, I could see that causing a fair amount of issues... and at least right now, the IOMD port of RISC OS 5 is 32-bit, IIRC.

bhtooefr on 24/5/09 10:23PM
[ Reply | Permalink | Report ]

Doubtful. But a 32 bit OS brings as many advantages as it does disadvantages. Many podules are fundamentally very simple devices that new drivers can be developed for, or old drivers ported should their firmware authors be contactable.

rjek on 25/5/09 12:12AM
[ Reply | Permalink | Report ]

"a 32 bit OS brings as many advantages as it does disadvantages"

???

I'm really struggling to think of *any* advantages for a RiscPC user, but there are lots of disadvantages.

helpful on 25/5/09 2:58AM
[ Reply | Permalink | Report ]

Look further back in the thread. There are performance, memory efficiency advantages, possibilities of future CPU upgrades, and a single source base to maintain.

rjek on 25/5/09 9:52AM
[ Reply | Permalink | Report ]

performance - a 32bit OS on a RiscPC will be slower memory efficiency - eh? cpu upgrades - for the RiscPC, really? single source - not an advantage to a user

helpful on 25/5/09 3:49PM
[ Reply | Permalink | Report ]

What makes you think it'll be slower? There are many technical things you can do when in 32 bit mode that aren't feasible in 26 bit mode. Such as shared libraries, which lead to better memory usage and better cache usage, and thus faster. Future processor upgrades for the RiscPC are possible, but let's face it, it'd be silly. And single source *is* an advantage to the user; it's less likely to get screwed up.

rjek on 25/5/09 5:46PM
[ Reply | Permalink | Report ]

So if a user has a 2 slice Risc PC with 4 podules fitted, this 32bit OS will also run all their 26bit code or a you going to write all the 32bit drivers for the array of podules that are around for a Risc PC?

sa110 on 25/5/09 10:49PM
[ Reply | Permalink | Report ]

Read the discussion. And in any case, any people actually using RiscPCs rather than more modern, 32 bit systems or emulated systems really need to find a new bit of hardware.

rjek on 25/5/09 11:46PM
[ Reply | Permalink | Report ]

So, I'm sat here with a Risc PC, you have so far failed to convince me of the need to move to a 32bit OS for my system. I see my podules no longer working, some ancient apps no longer working. I seen nothing but hassle. Whilst my Risc PC is still working, I see no reason to use one of the virtual solutions or move to newer hardware.

sa110 on 26/5/09 8:04AM
[ Reply | Permalink | Report ]

That's because you're ignoring everything I've said in order to have an argument, no doubt. As for the old software argument; I've said it before, but not in this thread, but I'll say it again; the kind of people using two-decade old software when more modern alternatives are availble are so stuck in their ways, you can't expect them to upgrade anyway

rjek on 26/5/09 9:49AM
[ Reply | Permalink | Report ]

"the kind of people using two-decade old software when more modern alternatives are availble are so stuck in their ways, you can't expect them to upgrade anyway"

Broadly, I agree with this. However, let us assume the following: -

Risc PC Owner, Running RISC OS 6 Running Podules such as a SCSI Card and a TV playing one Perhaps the user has both modern version of some software but then also has lots of old utilities which are used on a regular basis.

If ROL suddenly decided that all the next release of their version will be 32bit only, then why as a user of the above hypothetical system would I want to upgrade it if large number of apps and my podules will no longer work?

sa110 on 26/5/09 10:43AM
[ Reply | Permalink | Report ]

Who said they'd want to upgrade? I'm ignoring RiscPCs, as should everybody else. They're painfully ancient. And all current emulation is targeted on RiscPC-like hardware due to its simplicity, where the podule problem doesn't exist.

rjek on 26/5/09 12:32PM
[ Reply | Permalink | Report ]

What exactly are you trying to argue anyway? A 32-bit OS is pointless, so ROOL should stop?

Whatever you are trying to argue, you surely cannot be saying that RISC OS needs more support when you're sat there with hardware which is around two decades old and insisting that there is no need to upgrade any of it.

adh1003 on 26/5/09 10:08AM
[ Reply | Permalink | Report ]

"What exactly are you trying to argue anyway? A 32-bit OS is pointless, so ROOL should stop?"

The discussion on here at the moment is more about the value of a 32bit OS on a Risc PC, this could be either 32bit RISC OS 5 or 32bit RISC OS 6. At no point has anyone said, including myself that a 32bit OS is pointless. Without the 32bit OS would not have had either the Iyonix or the A9home.

"Whatever you are trying to argue, you surely cannot be saying that RISC OS needs more support when you're sat there with hardware which is around two decades old and insisting that there is no need to upgrade any of it. "

I have simply been pointing out that users may not want a 32bit OS on their Risc PC because it would inevitably stop many if not most of their installed podules working not to mention any old software/utilities they use.

On the other hand, if the Risc PC user does not have any podules, and only uses modern 32bit ready software, then a 32bit version of RISC OS (5 or 6) may just be what they are looking for.

sa110 on 26/5/09 11:25AM
[ Reply | Permalink | Report ]

Is RO5 32 bit or is it 32 bit neutral? If the latter then could it not be used in either mode on a Risc PC? Depending on whether 26 bit support is needed. (Maybe something similar to aemulor could even switch modes as needed.)

jess on 31/5/09 2:40PM
[ Reply | Permalink | Report ]

And so reads the epitaph of RISC OS.

So many people won't stop using their ancient old hardware and software, that there is no chance of new hardware ever being commercially viable, no chance of really modernising the OS, leaving RISC OS incapable doing the things users of every other platform take for granted.

druck on 26/5/09 10:50AM
[ Reply | Permalink | Report ]

"And so reads the epitaph of RISC OS"

...a decade ago. Since then it hasn't really had any significant commercial viability IMHO so it long since became a hobbyist OS and if anything, this is more or less proved by the very small scale operations of the likes of Castle and ROL. As such, it can potentially actually move along in directions which a commercial company might not have taken because of commercial pressures.

Whether there are enough people left in the community with the skills, time and willpower to make any of those changes remains to be seen, though we've certainly seen a big surge in interest at ROOL since we started making real ROMs. But even a "big surge" doesn't get you all that far since it's all relative to the level of active developer interest beforehand, which was broadly speaking "almost nothing" (unsurprisingly).

There are still one or two killer applications around - ArtWorks being an obvious example - but finding anything unique is difficult these days, unsurprisingly, while finding missing stuff is easy. I do worry that a lot of the directions people want to take the OS are being considered because they want to make porting easy, but if all you want to do is port applications from Linux, you may as well just go and use Linux instead! You'll have cheap, powerful hardware available and a wide range of distro choices. Some of ROL's developments seem more aimed at making life a bit easier for the applications programmer and that seems to me to be a more astute direction to take.

I always found that using Windows was just awful in comparison to RISC OS and - now, I know I'm going to ruffle feathers here but I'll say it anyway - Linux GUI environments just seem to be a bit of a bad joke. Bad experiences with Mac OS System 7 through 9 along with its very poor and buggy initial implementation as OS X had meant I dismissed and ignored Macs for many years, but once OS X reached v10.4 (Tiger) it became an architecturally complete, interesting and viable alternative which I'm very happy with as an ex-RISC OS user. While it has its fair share of faults, so does RISC OS. I just don't use my Risc PC or Iyonix for anything other than hobby development or ROOL work anymore.

I still wish operating systems routinely came with a "send to back" button and an option to avoid forcing a window to the front of the stack when focused! At least on Linux, RISC OS and OS X you can scroll windows which aren't frontmost (unlike on Windows) and on Linux and OS X you can move windows even when the application is busy or using blocking modal dialogue boxes (unlike Windows and RISC OS).

So where do I think this leaves RISC OS? For fun; for a hobby; for those few applications you have which you still really like. And ultimately, given a free OS from ROOL and a free emulator in the form of RPCEmu, what's so bad about that?

adh1003 on 26/5/09 11:18AM
[ Reply | Permalink | Report ]

Your post nicely sums up (to me) what I've been saying for a while now; there's no money in RISC OS. Let's just amuse ourselves and not delude ourselves with it.

rjek on 26/5/09 12:38PM
[ Reply | Permalink | Report ]

"And ultimately, given a free OS from ROOL and a free emulator in the form of RPCEmu, what's so bad about that?"

If I were given a choice over a 26bit ROOL RO5 and a 32bit ROOL RO5 for use with RPCEmu (or a real RPC for that matter) I would pick the 26bit, at least until more of the applications that exclusively use the 32bit versions features are released.

Which probably means I'd being using 26bit for ever ...

flibble on 26/5/09 2:36PM
[ Reply | Permalink | Report ]

The problem here is that to make good use of some of the 32 bit features brings problems for 26 bit users, so there's a bit of a boot-strapping issue, cause by people being stuck in the past. Although if they weren't, I don't suppose we'd be here at all

rjek on 26/5/09 3:11PM
[ Reply | Permalink | Report ]

To what extent can Aemulor bridge the gap (slow or otherwise)?

adh1003 on 26/5/09 6:08PM
[ Reply | Permalink | Report ]

I've used Aemulor from the beginning. Apart from a few 28M limit problems it has been trouble free and virtually invisible in operation.

Here is the complete list of what I use it for. Publish+ and Tablemate I would class as essential; EFFCAT and SuperGram I would miss; the rest I could manage without.

!Chords !Meteors !ColourMap !MineHunt !CrossStar !PDCAT !D2Font !PREMIER !Dilo !Publish+ !Dis-le !smbserver !Dis-leII !Snapper !Drummer !SuperGram !Dungeon !TableMate !EFFCAT !TemplEd !EFFmanual !Trace !effTTT !TransDDF !Equasor !TransFSI !Fresco !TransIMP !GraphDraw !TransPCD !ImageFS !TransRTF !LinkCheck !TranTIFF+ !WordWorks

JohnR on 29/5/09 3:18PM
[ Reply | Permalink | Report ]

I am pretty sure that all EFF software was converted to being 26/32bit neutral quite a long time ago.

hubersn on 30/5/09 9:03PM
[ Reply | Permalink | Report ]

and TemplEd was replaced by the much improved WinEd.

riscosopen on 31/5/09 12:08AM
[ Reply | Permalink | Report ]

GraphDraw, Trace and Snapper (I guess you refer to David Pilling's software) were all 32bitted a long time ago, as well as smbserver. Equasor should have been replaced long ago by Formulix, which also has a free 32bit upgrade.

So it is basically Impression and Tablemate. Hopefully, Tablemate will be 32bitted soon. And Impression is best replaced by Ovation Pro. Which pretty much eliminates the rest of your list.

hubersn on 31/5/09 8:51PM
[ Reply | Permalink | Report ]

"Which probably means I'd being using 26bit for ever ... "

Or until you make the move to a 32bit machine such as the Iyonix and A9home or possibly the beagleboard when the OS is finished. If you then still need to use your old software, you could purchase Aemulor.

Personally, I can't say I need to use any old software on a regular basis. However these days there are other means such as RPCEmu.

sa110 on 29/5/09 12:55AM
[ Reply | Permalink | Report ]

I would stick with adjust on my Risc PC because it is better, but I'm unlikely to be buying a new version of RISC OS that only runs on my secondary machine.

It would be nice if the HAL could be a seperate component that could be supplied with a machine and then either free RO 5 or bought RO 6 be loaded.

(As I understand it currently the HAL is only a separate entity at the source code level.)

Obviously this would require co-ordination and a a lot of work for ROL and ROOL.

jess on 22/5/09 11:45AM
[ Reply | Permalink | Report ]

As has been discussed before, distributing RISC OS and HAL separately is a non-starter. Additionally, the ROL and CTL branches have likely diverged so much, it would be next to impossible to get them to interwork, even if you could get the companies to co-operate on making an OS people hold dear survive, rather than trying to make money out of it.

rjek on 22/5/09 5:47PM
[ Reply | Permalink | Report ]

"5. Select developments

Overall, if we ignore the long gap between Select 3 and Select 4, the Select scheme has been a good success. Both RiscPC and Virtual RiscPC users have enjoyed regular updates to the OS and its ancillary applications. The question is how far should such a scheme go?

I think RISCOS Ltd has missed a trick here - after the fuss over ROL producing a version of Select for the Iyonix, the Select scheme should progress from being all about the operating system to being a product for the platform as a whole.

Would it not be feasible for Select development to be split, say six months on operating system work and six months on applications for the platform - such as finishing off the port of Firefox, porting the Gnash Flash player or an open source Java engine?

These are only examples. I’m sure other people would have their own ideas for general developments that would benefit the platform as whole. Such activity would almost certainly make users of RISC OS 4, 5 and 6 consider subscribing to the scheme and thus channelling additional funds into new developments. Sadly, it's unclear just how much in the way of extra funds this may bring or how effective this split development approach will be - but I'm sure it's worth considering."

Veering wildly back on topic...

I couldn't agree more with this. rjek and others are arguing about the worth of various features of Riscos6 (see above), and I have to say I think both sides of the debate miss the greater point raised in Paul's article.

According to personal preference SunRaster import may well be mostly useless while image file thumbnails may well be mostly useful---but if some time were spent working on applications at the expense of working on OS polish then that would be vastly more appreciated than either.

Monty on 22/5/09 9:32PM
[ Reply | Permalink | Report ]

"Would it not be feasible for Select development to be split, say six months on operating system work and six months on applications for the platform - such as finishing off the port of Firefox, porting the Gnash Flash player or an open source Java engine?"

Unfortunately, all of these need huge OS improvements to be truly usable. If RISCOS Ltd. wanted to make a contribution to the application world, I look forward to their efforts improving NetSurf, ArtWorks and Impression.

rjek on 25/5/09 11:48PM
[ Reply | Permalink | Report ]

And Cineroma? What *actually* happened to that?

Monty on 26/5/09 1:08AM
[ Reply | Permalink | Report ]

Sorry, Cinerama. I think.

Monty on 26/5/09 1:14AM
[ Reply | Permalink | Report ]

and Fresco was replaced by NetSurf.

IvanDobski on 31/5/09 1:13PM
[ Reply | Permalink | Report ]

There are a number of points that have been made here and I'm going to deal with them in one message rather than replying to each individually.

Druck - you seem to have a very selective memory. You were a Select subscriber for 2001 and 2002. The ARM Club then paid for your subscription in 2003 and 2004. So you weren't a subscriber during the period that the 32 bit work was being done. The ARM Club were subscribers to the scheme for a further year in 2007, but you yourself haven't paid for a subscription since 2002.

VinceH, you state that you can't test your software if you were to use a Select feature. Is that because you don't have Select? I get the impression from a number of the similar comments from other people as well that you think that developers should get free access to copies of Select in order to test their software. There used to be around 1,200 registered developers with Acorn. Is it your submission that if we gave away free copies to developers then they would use Select features? The only problem here is that we would have everyone claiming to be a developer.

not_ginger_matt - I guess your are also not a Select subscriber. I am also confused as to why you are complaining that you had to re-implement a feature that Select has? Shouldn't your anger be directed at ROOL for not having a feature that you wanted in RISC OS 5? Your argument seems to run that unless a new feature is available to every version of RISC OS, then either developers shouldn't use it, or else they should spend the time re-inventing the wheel. My view is that unless users keep up to date with the latest OS then it will stagnate and die. That is a commercial reality. 64K screen modes are a benefit to users, and developers should try and support new OS features in their applications.

Consider this scenario. If there was no RISC OS Six, would you have held back on producing tinct and waited until RISC OS 5 had it, or would you still have gone ahead and written it anyway so that RISC OS 3.x users could use NetSurf?

Rjek. If every new feature was made freely available to all previous users, then where exactly is the income to pay for the development to come from? If users want the latest features then they need to have the latest version of the OS. Everyone gains in that scenario. If developers don't use new OS features because not enough people will be able to use it, then no new software will ever be developed.

riscosboss on 2/6/09 11:57PM
[ Reply | Permalink | Report ]

You've once again completely missed the point.

Tinct (an alpha-blended sprite plotting module) was written because the Select alpha plotting wasn't made available to non-Select users. This meant that NetSurf couldn't render PNGs correctly and an alternative solution was required.

Flash forwards 5 years and various other applications also use Tinct due to the same lunacy that NetSurf faced. Except now, RISC OS Six supports a new 64k screen mode. A mode that Tinct, and thus any application that uses it, doesn't support.

Do you see the stupidity here? RISC OS Ltd's lack of foresight has created itself a dependency on a module that provides functionality that it already has. Given that the source code for Tinct isn't openly available, imagine if I'd turned around and had the same attitude as RISC OS Ltd., deemed the updates non-commercially viable and left the 64k screen modes somewhat of a dead duck for anyone wanting to use !NetSurf.

Also, such low-level stuff as this really can't be tested for and the used accordingly. For a start you'd still need to write the fall-back code anyway, so all you get is twice the code base to maintain. I'd also suggest that a simple look around will show that RISC OS Ltd. has no idea how to stop the OS stagnating and dying.

Regarding RISC OS 5's lack of such API support, it sits precisely alongside the support for 3.5 and later to me. Why couldn't RISC OS Ltd. have worked with whatever relevent parties and made it happen?

not_ginger_matt on 3/6/09 12:56AM
[ Reply | Permalink | Report ]

"Also, such low-level stuff as this really can't be tested for and the used accordingly. For a start you'd still need to write the fall-back code anyway, so all you get is twice the code base to maintain. I'd also suggest that a simple look around will show that RISC OS Ltd. has no idea how to stop the OS stagnating and dying."

There's no point getting technical here; it's clear that Paul's either unable or unwilling to understand the issue.

rjek on 3/6/09 6:53AM
[ Reply | Permalink | Report ]

"Rjek. If every new feature was made freely available to all previous users, then where exactly is the income to pay for the development to come from? If users want the latest features then they need to have the latest version of the OS. Everyone gains in that scenario. If developers don't use new OS features because not enough people will be able to use it, then no new software will ever be developed."

Are you being intentionally obtuse? I've explictly said that user-facing features shouldn't be; only APIs should be if you actually expect software to use it.

Try reading for a change, rather being deliberately argumentative to make you look powerful to the last few remaining people who think you're relevant.

rjek on 3/6/09 6:52AM
[ Reply | Permalink | Report ]

Isn't part of the issue that many of the developers that are left are (for whatever reason) not subscribers to the Select scheme. With this being the case, how do you expect them to support features of Select, if they do not have access to it? How exactly you get around this, I do not know.

sa110 on 3/6/09 7:37AM
[ Reply | Permalink | Report ]

"Druck - you seem to have a very selective memory. You were..."

Seriously Paul, you really need to stop publicly posting information about what people have and have not bought from your company! I'm pretty sure there are laws that forbid you to disclose such information.

Gulli on 3/6/09 8:34AM
[ Reply | Permalink | Report ]

There certainly are and from my understanding of them, they've been broken. I suggest it would be wise for Paul to stop posting in public forums given as he seems unable to stop this sort of very personal customer information from creeping into most of his posts.

While it would be a pity not to be able to discuss with a representative of ROL, if Paul is unwilling to maintain either a civil and none- vindictive tone or a sense of dignity towards the privacy of his own customers, then it may be unavoidable.

Monty on 3/6/09 10:45PM
[ Reply | Permalink | Report ]

riscosboss:

"VinceH, you state that you can't test your software if you were to use a Select feature. Is that because you don't have Select?"

That's correct, I don't - and there's a very good reason for that: My development machine is an Iyonix, and I don't have the space to have another machine alongside it and my laptop (the latter being my primary machine for all other work).

Installing VA on the laptop could arguably be a solution, but I did that on my previous (previous) laptop and, quite frankly, I hated using it with a vengeance - so that's a complete non-starter. I might revert to emulation again when the fat lady not only sings, but smashes all the glasses in the process - but even then, I might have to give serious consideration to just how much I disliked it before deciding.

"I get the impression from a number of the similar comments from other people as well that you think that developers should get free access to copies of Select in order to test their software."

That's not actually what was said, though, is it? Being generous, you appear to have misunderstood or misread a number of comments made here. Being less generous, you could be misrepresenting what people say in order to make your own angle stronger.

Given that the people in question seem to be the developers - the very people who you want to use Select features in their software - I'd hope it's misunderstanding, because it would surely be incredibly daft to misrepresent the people who you should be trying to get on side.

But that's an aside. Getting back on topic...

As has been said more than once, you should give serious consideration to making API improvements, rather than user-facing improvements, available outwith Select subscriptions. This would surely be beneficial to the platform as a whole, chiefly because it will enable developers to better support those APIs - and at the day's end, developers are vital to the platform.

However, having said all of that there is still the flaw of the computers in use. I fuly reserve the right to be wrong, but I get the very distinct impression that most developers use Iyonix PCs. At the very least, therefore, you need to find away of releasing components including new APIs for that platform if you want developers to use those new APIs, IMO.

VinceH on 3/6/09 9:42AM
[ Reply | Permalink | Report ]

Paul, you claim I have selective memory, but it was you who agreed we could switch my personal subscription in to the name of the club (of which I am I chairman), so we could use Select on the club's computer at shows, in order to PROMOTE YOUR PRODUCTS, and this is the thanks we get.

But anyway that has nothing to do your current decision to condemn the good work that has gone in to design of new APIs as an irrelevance, to not only the majority of RISC OS users who aren't running Select, but also to the future of the platform, as your OS variant doesn't run on the new hardware which will be the basis of the next generation of RISC OS devices.

It's time to wake up and get some business sense, you live or die by the success of the platform as a whole, and can't survive hiding in your ever shrinking corner.

druck on 3/6/09 10:42AM
[ Reply | Permalink | Report ]

The only was that I can see Select having a future is if it can be loaded onto machines that run Open as an upgrade.

One way to achieve this would be to ensure Open and Select always have similar enough internal structure so that a pick and mix approach could be used.

Then it would be worthwhile for Iyonix (and future machines) owners to join select.

I have recently started using my Adjust Risc PC again as a second machine. It certainly is a lot nicer OS than on my Iyonix. Unfortunately, it is too slow for many things I can just about get away with on the Iyonix, and is just used as an RDP terminal for a lot of tasks.

With respect to the Image rendering system, ROL really need to produce a freely distributable module that provides the API to other versions of the OS. (Even if this module doesn't support all the formats the full version does, eg just Sprites and any 3rd party add-ons). Without this, the whole system is totally pointless. And that would be a shame.

jess on 3/6/09 1:11PM
[ Reply | Permalink | Report ]

> One way to achieve this would be to ensure Open and Select always have similar enough internal structure so that a pick and mix approach could be used.

Shudder. Support nightmare!!

riscosopen on 3/6/09 5:13PM
[ Reply | Permalink | Report ]

I mean using modules, not mixing source code.

jess on 4/6/09 12:51PM
[ Reply | Permalink | Report ]

Even so, there'd be an almost infinite number of possible combinations, if each module could come from more than once source.

Stoppers on 4/6/09 2:04PM
[ Reply | Permalink | Report ]

but isn't that already the situation? And In reality there are only likely to be a few combinations anyway (especially if the upgrades are packaged.)

Select Select plus all bits of open that are better. Open Open plus free modules from ROL Open plus all good bits from select.

The current situation is that either OS could have any or all of the suitable modules supplied by the other company (clibrary, toolbox, dosfs etc.)

jess on 4/6/09 3:44PM
[ Reply | Permalink | Report ]

No, it's not already the situation. The APIs that many of the modules use to talk to each other have diverged. What people manage to run from ROOL on ROL is by luck, rather than design.

rjek on 4/6/09 4:19PM
[ Reply | Permalink | Report ]

That's something of an exaggeration - most things will still work perfectly well. After all, there hasn't been a vast amount of development in either branch in terms of updating existing modules and only a fraction of that development will have been divergent to such an extent as to cause failure.

However, the main problem is in all of those odd little edge cases. ROL would be mad to endorse a pick-and-mix solution because where would their support obligation to the user end?

And as the ROOL staff work for free, our support position would be no different to today: you get what you're given.

riscosopen on 4/6/09 8:53PM
[ Reply | Permalink | Report ]

I was thinking in terms of the pick & mix being done by ROL (perhaps as a riscpkg). ROL could contribute developement to any components they endorsed if there were issues. There would be two endorsed configurations. Select as is and Open enhanced.

Currently people will be doing their own pick and mix, leading to far more potential combinations.

jess on 5/6/09 12:46PM
[ Reply | Permalink | Report ]

Haven't dropped by Drobe in ages........ So glad to See Paul Middleton is so upeat about his sales. Hopefully that means ROL might soon pay Castle some of the several years back royalties they owe to Castle. I mean Paul, payment of royalties was part of the "spirit" of your original agreement with E14 was it not?

PeteWild on 28/7/09 1:55PM
[ 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

Internationalising RISC OS
Unicode, i18n and more explained

30 comments, latest by caliston2 on 16/7/03 8:57PM. Published: 10 Jul 2003

Random article

School wins IT award with RISC OS
Upgrades to 4.39, uses helping hand from Apple Macs

8 comments, latest by rmac on 11/4/06 1:56AM. Published: 10 Apr 2006

Useful links

News and media:
Iconbar • MyRISCOS • ArcSite • RISCOScode • ANS • C.S.A.Announce • Archive • Qercus • RiscWorld • Drag'n'Drop • GAG-News

Top developers:
RISCOS Ltd • RISC OS Open • MW Software • R-Comp • Advantage Six • VirtualAcorn

Dealers:
CJE Micros • APDL • Castle • a4 • X-Ample • Liquid Silicon • Webmonster

Usergroups:
WROCC • RONE • NKACC • IRUG • SASAUG • ROUGOL • RONWUG • MUG • WAUG • GAG • RISCOS.be

Useful:
RISCOS.org.uk • RISCOS.org • RISCOS.info • Filebase • Chris Why's Acorn/RISC OS collection • NetSurf

Non-RISC OS:
The Register • The Inquirer • Apple Insider • BBC News • Sky News • Google News • xkcd • diodesign

© 1999-2009 The Drobe Team. Some rights reserved, click here for more information
Powered by MiniDrobeCMS, based on J4U | Statistics
"I'm not sure I could have trusted you six months ago"
Page generated in 1.8886 seconds.