RISC OS 6 to power Select 4 | RISC OS news and information from Drobe


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

RISC OS 6 to power Select 4

Published: 18th Oct 2006, 17:11:37.

Free preview promised to hearten Select subs [Updated]

Select 4 will be powered by a 26/32bit neutral RISC OS 6, RISCOS Ltd announced today. The next heavily awaited installment of Select will come in two builds - a 26bit mode build for RiscPCs and A7000-class machines, and a 32bit build for A9home computers. Drobe understands a free preview version running RISC OS 4.4x will be offered for Select subscribers to download, with a suitably bug fixed RISC OS 6 later rolled out to fully paid-up subscribers. The news was announced this afternoon in Glasgow during the first day of the RISCOS Ltd northern roadshow.

RISCOS Ltd have privately fumed that punters could see RISC OS 5 as being superior to RISC OS 4 purely from comparing the version numbers, and have stuck to using the Select and Adjust brand names to describe their operating system releases. However, as the internal kernel version number has rolled past 10.00 after hundreds of changes and bug fixes were committed to the source code, ROL have seen fit to nudge the OS version to six.

Glasgow roadshow notes
An X300 and an Aria cube were on display on the Castle stand, with a very flash widescreen display. The RISC OS 5.12 ROMs have apparently have proved to be a very popular download since their release. On the subject of a future Iyonix employing a 1.2GHz XScale processor, Castle wouldn't comment and stated the usual policy that they won't discuss such potential products. There's a feeling they're pretty interested in it though, but said that the chips weren't actually available yet - the 800MHz part is, though. Incidentally, RISC OS Open are reportedly working flat out to get the whole operating system out within the next 12 months. When a RISC OS 6 leaflet was thrust into a ROS Open staffer's hands, he said he was surprised, but added it was good to get RISC OS out onto people's computers, and it didn't matter if it was Castle or ROL doing it - it would all aid the market.

Liquid Silicon were at the show and said there had been a steady flow of people in during the day, having opened at around 3pm. Alan reported he was having trouble tracking down ESP with regards to the 32bit MIDI drivers, and the USB MIDI driver development is said to have gone cold.

Next up was RISCOS Ltd, which had a plethora of computers on display, including a real surprise - RISC OS 6. Paul Middleton admitted that it doesn't work on an Iyonix, but it does have a flashy blue loading screen and a 'six' in the cog switcher icon on the iconbar. It was running on a RiscPC, and an A9home. The desktop had a cool blue backdrop, and VirtualRiscPC was running on a Mac.

Chris Evans of CJE Micros was selling fast on his stand, and had some A9homes in stock plus an A9home hooked up to a network.

There was a steady trickle of people coming in the doors with around 12 or so the most at any one time.

The key features of RISC OS 6 include a highly modularised operating system to aid portability, stability and maintenance of the OS; legacy support components; and updated graphics, toolbox, programming library, networking, and desktop components. Teletext support, the VIDC driver, device interrupt handler, software-based graphics operations, the real-time clock, the mouse pointer, CMOS RAM support, and hardware timer support have been abstracted out of the kernel and into their own separate modules. Support for obsolete BBC Micro-related functionality has also been abstracted out. AIF and transient utility executable checking is included to help block badly written software from running on systems.

Overall, the bullet point features for Select 4 are the hardware abstraction; increased stability thanks to additional memory protection and the new debugging tools developed in-house by ROL; the ZeroConf networking system to make managing a network easy; the restructuring and tidy up of the operating system; and the 32 bit conversion making it suitable for ARM 9 powered systems. The abstraction includes some graphics acceleration in a Viewfinder driver module, which opens up extra large screen modes and is independent of John Kortink's VF driver code. A similar module also exists for the A9home SM501 graphics chip.

The acceleration enhances line and rectangle drawing, and block operations, reportedly pushing Draw far faster than ArtWorks. The Viewfinder driver also caches sprites on the card video RAM, accelerating rendering further. The memory management system has been updated so that dynamic areas can be used in DMA operations and mapped to physical memory directly, which is useful for driver development, and can be adapted to power a virtual memory system. The video drivers use physical dynamic areas to access their VRAM. A Video Guard module ensures users are not left without a display if a fault occurs with a graphics driver.

With its 32bit compatibility and hardware abstraction, the OS has been described as a distant cousin to earlier releases of RISC OS 4, and unrecognisable compared to the design of RISC OS 5.

The last release in the yearly Select subscription scheme was pressed in mid-2004. RISCOS Ltd recently apologised to customers for the delays, adding that a lot of development time has been spent on making RISC OS 4 compatible with 32bit ARM-compatible processors and introducing the hardware abstraction. Since the release of Select 3, AdvantageSix have bank rolled further development of RISC OS 4, which saw the release of 32bit ROS 4.42 for Ad6's A9 product line. The new version number, namely six, and AdvantageSix's name are a coincidence, Drobe understands.

ROL's Paul Middleton told subscribers this month that the company is aiming for an October release; this is unlikely to happen, although development and bug fixing continues, Drobe understands. An Iyonix port of Select is also unlikely to happen unless greater cooperation between ROL and Castle emerges. ROL intend to charge £210 for a 2006-2007 subscription to Select, which will include RISC OS 6.

The RISC OS 6 announcement comes following the publishing of the RISC OS Select technical documentation online. ROL have denied that the flurry of Select related news is due to the appearance of RISC OS Open.

Glasgow show photos from RISCOS Ltd stand
Leaflet page 1, page 2 (copy of the RISC OS 6 website)

	Available "soon"
	Copyright RISC OS Ltd, acknowledges old Acorn
	Select 4 sources directory? An alternative form of open RISC OS, we think. Note the new Paint toolbar on the left hand side
	Note VIDC driver DA mapping in video RAM
	This info box is a mock up, before you get carried away
	Sixth volume for sixth version
	A preview black and white copy

Rewritten at 23:20 after new information and photos came to light.
A pre-release hard disc archive is also expected to be released soon.

Update at 20:20 19/10/2006
The full changelogs for Select 4 are here.

Links
RISC OS 6 website
News moles: Glasgow show reporting by Andrew Brown and Michael Drake. Photography by Michael Drake.

Discussion

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

Its very nice to see all this activity, but I'd really like to see Select on my Iyonix. What about customers actively wanting Select on their machine?

ROL has a much better GUI and RISCOS 5 has unicode support and other goodies. What do us customers have to do to get you lot to work together (without the A blaiming B game) ???????

markee174 on 18/10/06 5:40PM [ Reply | Permalink | Report ]

I think this is a bad joke, calling it the "latest, most up to date" version, and then telling you it will only work on old hardware.

When it works fully with the Iyonix's HAL/hardware the statement would be more suitable.

ajb on 18/10/06 5:57PM [ Reply | Permalink | Report ]

In reply to ajb: RO6/Select 4 is the latest and most upto date version of RISC OS, as developed by ROL. Their move to call it RO6 is in line with it now being 32bit and able to run on all RO4 Machines, from A7000 to A9home.

As they have already stated in the past and is well known, they have limited resources and will be releasing RO6/Select 4 for all of their currently supported computers before producing a version for the Iyonix(assuming they are able to obtain the relevant information from CTL).

sa110 on 18/10/06 6:04PM [ Reply | Permalink | Report ]

ROL launch 32-bit RISC OS 6?

Oh? Eh? So What?

:-)

)

Ike on 18/10/06 6:14PM [ Reply | Permalink | Report ]

sa110:

From your reply above, I assume that you agree that their statement, "latest, most up to date", should have been more conditional.

ajb on 18/10/06 6:18PM [ Reply | Permalink | Report ]

Are they even allowed to call it ROS 6 ? This is getting ridiculous. If ROS Ltd hAD 32 Bitted the os in 1999 like it was supposed none of this nonsense would have happened. ROS are the only ones to blame for the demise of the market and know they are driving the split even further. I personally hope that Castle releases ROS 5 roms or some other way to use RO Open on classic hardware!

highlandcattle on 18/10/06 6:21PM [ Reply | Permalink | Report ]

ajB:

It is justifiably the latest in some features but not in others. That is what is so infuriating :-(

markee174 on 18/10/06 6:30PM [ Reply | Permalink | Report ]

Well, what's going on here.

ROL making 2 positive announcements and Qercus back on sale..Que alternative universe music.

Seriously this is great news in that it looks as if the market is moving ahead again and ROL should be congratulated for the recent positive moves. What they now need to do is deliver the final bit of all this and deliver RO6 for the Iyonix before moving forward with merging the best bits of RO5 in with RO6. I also agree that the ROOL situation may have some bearing on all this.

As a Select subscriber and Iyonix user I'm feeling more positive than for a long time but I don't want to be part of RO5/RO6 marketplace struggle as that is the road to oblivion and if we are not careful a betamax/vhs/blueray/hdd type battle with RISCOS variants could be that road . As I have said before if ROL/ROOL/Castle work this one through then the goal of a combined and developing product available to all can be achieved.

In the meantime well done to all at ROL and I look forward to RO6/Select4.

bluenose on 18/10/06 6:34PM [ Reply | Permalink | Report ]

I think I just blew a fuse... WTF?!

Indeed, this is ridiculous. ROL should've released RO4.5 as their 'official' 26 / 32 bit neutral version. Although I've read the FAQ on their (somewhat good looking) RO Six webpage, I still find this move confusing. To me it seems ROL is playing a game.

So we now have:

RO3.7x - limited supported legacy RO4.0x - Standard version RO4.39 - RISC OS Adjust RO4.4x - A9home version RO5.xx - Iyonix version RO6.xx - Legacy & modern + possible Iyonix version.

Oh, and somewhere in between there's a Select scheme running. Amazing, isn't it?

hEgelia on 18/10/06 6:48PM [ Reply | Permalink | Report ]

in reply to Bluenose:

I agree, for once we have some good news. Lets celebrate that fact rather complain.

We should also celebrate the proffesional look of this and the PRM site.

Col1 on 18/10/06 6:58PM [ Reply | Permalink | Report ]

correction of typo: Should read..rather than complain.

Col1 on 18/10/06 7:01PM [ Reply | Permalink | Report ]

In Reply to hEgelia

Well if you strip out anything before 4.0x and take into account ROL's statement about 4.39 being the new base then it does start to make some sense if we users play a part and upgrade.

Essentially if I read it correct then Select4.39/Adjust 4.39 and A9Home are essentially the same, apart from some A9 specific bits.

RO6 would then be the same version what ever system it was on.

Ok this doesn't take into account the commitment to RO6 for Iyonix but it seems that ROL are now starting to deliver on what they said i.e

Deliver RISCOS4.4 for A9 Deliver Select 4 for legacy Deliver Select for A9 deliver Select for Iyonix

It seems that the steps for Select have now become a rebranding aka RISCOS6.

Seems logical to me , if of course my thinking is correct.

Still only a few days to go to hear it direct from ROL. But I do hope they have some sort of correlation chart to distribute as I'm sure just go on and on in the good old RISCOS tradition of claim and counter claim.

bluenose on 18/10/06 7:09PM [ Reply | Permalink | Report ]

Great news all round.

I can understand ROL being keener on customers like Ad6 and RPC users at the moment, rather than on having to work round the Iyonix compatibility issues without Castle's positive involvement. That's just being realistic.

OTOH I have a sneaking suspicion that making the Select PRMs free has dropped the revenue stream on their sales from effectively zero to, er, much the same really. Possibly more a case of nothing to lose than a genuine understanding of its importance, but it's a start. Let's hope it's rewarded by renewed activity.

Hopefully too, with the Select PRMs out there and RISC OS 5 becoming available to hackers, we might see a degree of OS re-convergence by adding Select/RO6 compatibility features to RO5. That has to be good for programmers, and hence also for users.

steelpillow on 18/10/06 7:10PM [ Reply | Permalink | Report ]

Excellent news that ROL have 'seen the light' and made a sensible decision on version numbering. I for one will put off buying RISC OS 5.12 until an official ROL 6.xx version is available from ROL for the Iyonix and then make a decision based on price and source tree compatability. Development of the OS should never have bifurcated (WUB) but this announcement should bring things back on line.

CKH2 on 18/10/06 7:15PM [ Reply | Permalink | Report ]

The thing I find confusing about all of this is ROL's pricing policy. If RO6 is going to include some but not all S4 features, and as they say "RISC OS Six is the base from which all future versions of RISC OS, to be released by RISCOS Ltd", then why don't they simply release RO6 as a stand alone product as an upgrade for all RO4 based computers, rather than insisting on £200 for new members to join the Select scheme to obtain? Select members(of which I am a member) would not be p***ed off, because they would still get the extra Select 4 features. Surely that would make more sense.

sa110 on 18/10/06 7:16PM [ Reply | Permalink | Report ]

CKH2:

Select is likely to run ontop of RISCOS 5, not replace it, so putting off buying 5 will just reduce the chance of developments there.

markee174 on 18/10/06 7:18PM [ Reply | Permalink | Report ]

I think at the RO roadshows this week, people need to harras Paul Middleton about RO6 for the Iyonix. The more harrasment he get's may help him make the final decision to actually get it done.

sa110 on 18/10/06 7:20PM [ Reply | Permalink | Report ]

I presume the A9home retail/reflash version will be 6.xx once Adjust3 is complete?

CKH2 on 18/10/06 7:20PM [ Reply | Permalink | Report ]

In reply to ajb: I suggest you go back to the RO Select site and actually read the first paragraph again. The first sentence says "RISC OS Six is the latest, most up to date, version of RISCOS to be developed by RISCOS Ltd". Which is exactly what I said in my posting. I don't see any problem with what they are saying about RO6 at all.

sa110 on 18/10/06 7:22PM [ Reply | Permalink | Report ]

And, for the fourth year, they are telling IYONIX users the following: "We are still planning on producing a version of RISC OS Select for Iyonix users. Iyonix users may contribute to the development of RISC OS Six by renewing their Select subscriptions. New Iyonix users who wish to join the Select scheme will have to pay the same initial subscription prices as those listed above for other users. "

Or, slightly paraphrased in my own words: "Give us your money, but we won't promise you get something for it." The only sensible advice for IYONIX users is to cancel their subscription as soon as possible.

And I guess those whose subscription ran out during June 2005 and July 2006 will be delighted to hear that they will be able to download a pre-release version for their subscription fee - the rest of their money presumably went into financing the A9 version of RISC OS.

hubersn on 18/10/06 7:24PM [ Reply | Permalink | Report ]

I see the anti ROL brigade are fuming again. There was announcement about open sourcing RO5 and it was all positive here on drobe. ROL announce two positives in one day and there's bitching and sniping.....why?

The OS fork developed by Castle ONLY works on iyonix. The OS fork developed by ROL works on all legacy machines and on the latest hardware which is exactly the situation that existed in the acorn world for years. So if any mud is to be slung do so in the direction of Castle.

This is good news for Christ's sake.

mripley on 18/10/06 7:28PM [ Reply | Permalink | Report ]

It's just a number, and with all RO version numbers and module numbers, it's still impossible to determine any idea of feature set from that number since the source code split.

Be the first with RISC OS 7 [link] [link]

flibble on 18/10/06 7:32PM [ Reply | Permalink | Report ]

In reply to fibble: Nice Lego A7000 case

sa110 on 18/10/06 7:39PM [ Reply | Permalink | Report ]

In reply to markee174:

Haven't Castle downsized their development team? That's more likely to hinder development than not buying RO 5.12....

CKH2 on 18/10/06 7:41PM [ Reply | Permalink | Report ]

CKH2:

So how will Castle hire developers if no-one buys the software upgrade?

markee174 on 18/10/06 7:43PM [ Reply | Permalink | Report ]

Why bother buyingth software upgrade, if you know there are no developers to continue providing the upgrades?

sa110 on 18/10/06 7:45PM [ Reply | Permalink | Report ]

in reply to hubersn:

I fully agree with your advice.

Ike on 18/10/06 7:45PM [ Reply | Permalink | Report ]

mripley:

Castle got quite a grilling from some quarters over their open sourcing RO, and we still don't know the exact terms of the license... The ROL fork only works on old hardware, emulation and the A9. It does not support the fastest RO machine.

markee174 on 18/10/06 7:49PM [ Reply | Permalink | Report ]

"The ROL fork only works on old hardware, emulation and the A9."

Or, to put it another way, the ROL software works on all current hardware except the Iyonix. The Castle fork only works on Iyonix.

More spin, anyone?

jms on 18/10/06 8:01PM [ Reply | Permalink | Report ]

I should have read the FAQ more carefully as I seem to have got confused on the relationship of RISCOS6 to select 4.

From the site:

"RISC OS Six is the base from which all future versions of RISC OS, to be released by RISCOS Ltd, will be developed. RISC OS Six Preview contains a number of features over and above those that are necessary solely to accommodate the 32 bit conversion of RISC OS 4.39 (Select3i4). However only Select subscribers will get the final release of Select 4."

I agree with mr ripley that this is positive news and as I said before lets not get into the Betamax/VHS type mode. ROOL/Castle/ROL/Qercus have all made some positives announcements over the past few weeks so lets rejoice in that instead of the more recent debates on the lack of news/developments etc.

bluenose on 18/10/06 8:17PM [ Reply | Permalink | Report ]

Has anyone found the minimum requirements for RISC OS 6 on the various hardware . . . ?

RickCB on 18/10/06 8:18PM [ Reply | Permalink | Report ]

Utter madness. Releasing a new version of the OS that officially only works on the RiscPC and the A7000. It should have been made to work officially on the Iyonix first. The Iyonix is the way forward, it was teh first machine to have a 32bit version of the OS and with its HAL broke the dependency on proprietorial Acorn designed components. Whether you like it or not the RiscPC is obsolete hardware, it is time to move on.

JWCR on 18/10/06 8:47PM [ Reply | Permalink | Report ]

And they have moved on. It'd called the A9...

tweety on 18/10/06 9:17PM [ Reply | Permalink | Report ]

tweety:

So I should abandon my 600mhz Iyonix and upgrade to a 400mhz A9?

markee174 on 18/10/06 9:22PM [ Reply | Permalink | Report ]

Yesterday's announcement of the publically available free documentation on Select was a good move by ROL.

This on the other hand I can't think of as anything other than "neutral".

How does renaming/rebranding something make it any better/worse? More likely just the same - but with a different splash screen and different message on an "*FX 0"

The article says "It will also be possible to produce A9home and Iyonix builds." But will that *actually* happen. It's a bit like saying "it is possible someday to have world peace", yep in principle sure - but will it happen?

The article said (and I quote) "For years, ROL have privately fumed that Castle's RISC OS 5 is considered wholly superior to their RISC OS 4 because of the version numbering". For the love of Pete, that's mad - utterly butterly bonkers. I refuse to believe ROL think that - people buy hardware based on technical considerations not because of the ordinality of some integer. If for marketing reasons ROL want to call a 32bit RISC OS 4 - RISC OS 6 that's their right - at the end of the day it doesn't really matter what it's called - it's what it *does* that counts.

AMS on 18/10/06 9:47PM [ Reply | Permalink | Report ]

ROL have said that like the Iyonix, an A9 build is possible, but not actually achieved yet. So no, they have not moved on. This is still a version of the OS for machines that no-one actually makes anymore.

JWCR on 18/10/06 10:02PM [ Reply | Permalink | Report ]

Looking at the leaflet from todays roadshow, unforunatelty it seems RISC OS Six is heavy on spin and light on new functionality. It doesn't contain many new user features (as opposed to internal changes) over and above Select 3i4, thats all *still* to come in Select 4.

Plus to get back on the Select scheme if you stopped subscribing ater the first year of zero releases, it seems to cost more than if you've never subscribed and signed up for the first time, which can't be right.

But anyway its no interest to me until a version is available to run on the computer I use, rather than the relic I use as a doorstop.

druck on 18/10/06 11:06PM [ Reply | Permalink | Report ]

> An Iyonix port is played down, with ROL prefering to "working with partners who actively want RISC OS Select features on their products."

...and that's where I completely leave the RISC OS market for good, folks.

I don't care whose fault it is. But the Castle/ROL situation is utterly, utterly stupid, and I want no more part in this ridiculous charade.

moss on 18/10/06 11:07PM [ Reply | Permalink | Report ]

in reply to moss:

the thought crossed my mind too. But having completely lost intrest in Select why should this regular attempts of ROL to provocate Iyonix users lead you to completely leave. Just ignore the OS for obsolete hardware and be happy :-)

Ike on 18/10/06 11:39PM [ Reply | Permalink | Report ]

The thing is, though - I'm not happy. I'm not happy with what RISC OS can do, when I can do far more on my Mac Mini.

But if Castle and ROL were co-operating, then we'd still have a chance - and perhaps the Acorn community would be more fun to hang around too. But they aren't, and it isn't. No, I mean it when I say - this is the moment when I have actually left RISC OS. This is the moment that has pushed me over the edge.

moss on 18/10/06 11:52PM [ Reply | Permalink | Report ]

I've been a critic of RISC OS Ltd. for a long time, but I currently feel they are really doing everything that's been asked of them (aside an Iyonix port) and more. Castle, on the other hand, seem to be doing pretty much nothing to fix long standing issues (unicode postscript printing for example.) Strangely the thing that guts me most at the moment is that it's the CTL version of RISC OS that's being open sourced. All the Select goodness will need to be replicated (which looks far less than a simple task from their documentation) and we'll end up wasting yet more effort to get to somewhere that should really be where we start from. All in all I truly applaud ROL for their recent announcements and I truly hope things work well for them.

not_ginger_matt on 19/10/06 00:41AM [ Reply | Permalink | Report ]

moss: I was thinking along similar lines, isn't this just another nail in the coffin for RISC OS?

True, everything appears to be rosy in ROL land but what has really changed? ROL finally releases the programming documentation that people have been wanting to get at freely for a few years already and they've upped the number after the OS name. They still haven't released anything but they've said they are going to release something soon, which isn't exactly news anymore.

But the real issue is that ROL and CTL have cemented the OS fork, there will never be a single OS fork unless someone buys both forks from the respective owners and joins them but the likelyhood of that is to understate a bit - very slim.

Gulli on 19/10/06 00:59AM [ Reply | Permalink | Report ]

In reply to quite a few:

1. I agree that IXONIX users should quit Select ... if they didn't so so quite some time ago since by now they (would have) paid quite a bit of money getting nothing - not even a guarantee that they will get something at all!

2. It seems that the next release of RISC OS from ROL (I mean the release of Select 4 which now seems to have been renamed) is the most expensive RISC OS release ever and it does amaze me that there seem to be quite a few Select subscribers still continuing to pay regularly.

3. So instead of finishing off Select 4 it seems that ROL prefers to spend time on other things like making the Select docs publically available (which is a good move), renaming their alleged product (I write alleged since it still has to be made available and until it is really released I fail to believe it will), redesigning the odd logo etc. for it, remake the website to go with etc.

4. Looking at the new site's pricing info I see one thing that has changed: If you subscribe now you will be able to download the preview of the RISC OS Six, that is you are guaranteed at least one download. As usual the additional things will come by (or not) via the Select scheme.

5. It seems to be a weird world: Now - sorry: in the future - the higher RISC OS version number is for the very old hardware, or the newest one, or the fake (virtual) one. But what the heck, RISC OS users are used to strange version numberings with CLib and Toolbox modules already.

6. And as for the much-commented <RISC OS Six is the latest, most up to date, version of RISCOS to be developed by RISCOS Ltd.>: Since to the best of my knowledge RISC OS Six is not there yet this statement is interesting indeed :-)

Shouldn't it read <RISC OS Six will be...> ;-)

7. And as usual the ROL-CTL-should-work-together issue is not changed by this..

Anyhow it is good to see that something happens ...

hzn on 19/10/06 06:24AM [ Reply | Permalink | Report ]

Anyone who isn't happy with this announcement: Join up and fix the things that YOU think are not right? Why blast those who can, and do try to make life better?

I see this as a good move, AND wish I had the money to invest. I was thinking of starting a company..and buying A9Homes to run it from. (I hope 32 Bit Impression makes an appearance soon)

I am also trying (not not succeeding) to code (BASIC, Java, and XML is about it atm).

as I am say about the closure of the Hospitals...if your not with us, your against us.

www.savehemel.com

em2ac on 19/10/06 09:14AM [ Reply | Permalink | Report ]

em2ac: "Anyone who isn't happy with this announcement: Join up and fix the things that YOU think are not right?"

Since when have Select subscribers had any influence over ROL? They are treated like complete fools, and expected to keep paying year after year regardless of the lack of development. Users of old machines at least have a free bugfix to follow on from Select 3i4 and the amost beleivable prospect of Select 4 only another promise and subscription fee or two away. However Iyonix users have to be twice the fools, told they have to continue to pay every year, when even if something is released it wont work on their machines, and with only the vaguest non commital suggestion that ROL might eventually get round to supporting their machine.

How much will we have paid for a new version of the OS by then? As much as the Iyonix itself? As much as 2 MAC minis? As much as 3 PC laptops? We might be foolish to stick with RISC OS these days, but we aren't complete idiots.

druck on 19/10/06 09:45AM [ Reply | Permalink | Report ]

In reply to em2ac: "Anyone who isn't happy with this announcement: Join up and fix the things that YOU think are not right?"

Join what? I assume you mean the group of non-Select-subscribers and ex-Select-subscribers...

hzn on 19/10/06 10:41AM [ Reply | Permalink | Report ]

David Ruck & Herbert Zur:

Actually i was meaning, join ROL...as a volunteer, sorry for that mis-understanding.

And yes, I am a fool as well, but I love the OS..

I indeed brought Adjust, and have been more-than-happy with it. :@)

It currently acts as my WebServer, DTP Machine, Technical Drawing Machine, NAS via Samba, and until recently was plugged into Telephonetics' Network, producing PDFs of network diagrams, Technical / Simple User Manuals, and helped with Network admin via RDesktop clients, (whilst my Dell Laptop was starting up).

and so when RISC OS 6 comes out as a ROM, my 2 SAs, and ARM 7 RiscPCs will be getting the upgrades (adjust to second SA RPC, RO 4 to ARM 7 RPC)

P.S. No offence meant by that last comment, just that getting annoyed at those who help, bugs me. Its like, "What's the point in signing this petition, it won't save the hospital ";@(

em2ac on 19/10/06 11:06AM [ Reply | Permalink | Report ]

Me:

Part time volunteer of course :@P

em2ac on 19/10/06 11:07AM [ Reply | Permalink | Report ]

Correct me if I am wrong, but didn't ROL have a license to RISC OS 4 and to its further development.

Has Castle now licensed ROL for RISC OS 6 ?

The Guildford show this weekend should be very interesting

ajb on 19/10/06 11:38AM [ Reply | Permalink | Report ]

I can already see Paul Middleton reading these reactions on Drobe and shaking his head. Indeed Paul, you can't win with words anymore. You've basically wasted all your credibility by taking people's money for years and replacing them with empty promises. There is only one solution in my opinion - organise an event, where ROL will deliver a product for users to take home. Preferably two products; RISC OS 6 and Select 4.

To ROL's credit, I have to say that, although I have mixed feelings about the 'SIX' numbering, it seems that this upcoming new version of RISC OS will be significantly enhanced with loads of bug fixes, optimisations and streamlining while being very well suited to base future developments for future hardware on. However, when most of us sit down in front of it we probably won't notice this. We'd find it may boot slightly faster or has slightly better network functionality. In all basically the same experience as v4.39.

I guess the reason ROL has chosen this is quite clear - they wanted to ensure their future, meaning they had to convert their RO sources to 26 / 32 bit neutrality. While working to achieve this, it was logical to make use of the low level work to fix bugs on a massive scale, streamline and optimise stuff, get rid of ancient and other useless stuff, abstract away / further modularize, etc. So, in the end, it may well represent the best RO has ever been and deserves credit for that. Right, but now it's time to stop gazing into and preparing for a future that may never come and start realizing there are a lot less people in the room.

In other words - I believe ROL has played their last 'Jack' card and have no more room for excuses. I know this has been said by others, as well as myself before, but I'm sure the thing moss said in an earlier comment has passed through the minds of others as well. It has with me. Some will find this the last straw and leave, whereas others experience this as a clear sign towards a more active and secure future. I'm still having mixed feelings.... I guess amid the various directions RISC OS has gone, the market has fallen apart, users have stopped bothering to keep themselves informed, resources have been wasted, etc. RISC OS Six could be the excellent new foundation it's intended to be. But for that to happen, it needs to run equally well on all machines and, hopefully before long, on an Iyonix II featuring promising new hardware. At least I can say, if that would happen it would surely restore a lot of faith in my heart.

By the way, just to make some things clear : RISC OS Six would run everywhere, except the Iyonix which will make it useless to a significant amount of users and developers. When it arrives, it would run on the A9home, though I'm not so sure exactly how different RO6 will be from the existing RO4.42 - I guess not by much. In any way, for anyone to use Select 4, RO6 is a prerequisite. Thankfully, when you get Select 4, you'd already have RO6. Theoretically. This all being. Finally, I'm quite curious how this all poses to developers - will they need to cater for widely differing OS versions, should they want an application to run on all machine / RO combinations? If so, and a developer will basically need to write several 'versions' of an application for it to function with all systems, this will be devastating for the market.

hEgelia on 19/10/06 11:50AM [ Reply | Permalink | Report ]

em2ac: If you treat ROL like a charity, they're treat you like a donor and not a customer. Donors don't receive products.

ajb: repeat after me "It's just a number".

flibble on 19/10/06 11:54AM [ Reply | Permalink | Report ]

In reply to fibble:

"I am not a number. I have a name !"

sa110 on 19/10/06 11:57AM [ Reply | Permalink | Report ]

em2ac: I was involved with beta testing for ROL for years, back when there was somerthing to beta test, which was great if you wanted to talk to the developers about technical aspects, but useless to address PM about strategy, he has always been off in his own disjoint universe. But if I was going to volunteer for anything now, it would be for ROOL open source effort (hold on, I have actually volunteered already), at least I then don't have to pay Paul Middleton to get the fruits of my own work.

ajb: ROL aren't concerned about licenses, you just cut out all the words and rearrange them in whatever order suits messers PM and AT, or at least thats thats how it works in the csa.misc amusement park. But then whos going to sue? The RISC OS market doesn't have enough money left even to get a good lawyer to answer the phone, and even if it did what would be the point in wasting it.

druck on 19/10/06 12:03AM [ Reply | Permalink | Report ]

"I am not a number, I am a free man"

flibble on 19/10/06 12:03AM [ Reply | Permalink | Report ]

Judging from the moderation scores, there appear to be lots (tens) of disgruntled Iyonix users out there who, admittedly, have received nothing yet from ROL. Perhaps it is CTL who need to cooperate more to move this forward...

CKH2 on 19/10/06 12:24AM [ Reply | Permalink | Report ]

What I would really love to see (apart from Castle and ROL announce they are merging their code which seems unlikely) is for them to at least get together and organise a coding weekend where they find a cheap central venue with a travel Lodge nearby and organise a programme of both groups telling us about their versions and also sessions on how to write code to run on multiple versions. It would also allow all the developers to get together and spark discussion and interaction.

markee174 on 19/10/06 1:12PM [ Reply | Permalink | Report ]

What I'd like to see is them locked in a room until they come up with an acceptable solution to this frankly ridiculous situation. Castle don't seem to show the slightest bit of interest in anything non-Iyonix, and I completely fail to understand ROL's attitude - do they want to sell copies of RISC OS or not? There's a market of Iyonix users, and they are stubbornly refusing to try to exploit it. Even if they just broke even with the development work required to get it working on the Iyonix the improvement in their reputation would probably more than pay off in the long run.

SimonC on 19/10/06 1:26PM [ Reply | Permalink | Report ]

Well, looking at Wikipeda for "Foundation" (as in RISC OS Foundation) you find the odd definition depending on the context. As for the non-name definitions the one for architecture doesn't quite fit leaving us with the charity one, i.e. "a kind of philanthropic organization, set up as a legal entity either by individuals or institutions, with the purpose of distributing grants to support causes in line with the goals of the foundation".

hzn on 19/10/06 1:33PM [ Reply | Permalink | Report ]

Druck:

Good for you man! :@)

keep up the good work! (DiscKnight helped me on many occassions)

em2ac on 19/10/06 1:35PM [ Reply | Permalink | Report ]

In reply to SimonC: "I completely fail to understand ROL's attitude - do they want to sell copies of RISC OS or not?" I'm not sure about that - to me it seems they want to sell Select subscriptions...

hzn on 19/10/06 1:35PM [ Reply | Permalink | Report ]

In reply to fibble:

I didn't quite think I got the last bit right

sa110 on 19/10/06 1:40PM [ Reply | Permalink | Report ]

Andrew Banks wrote>"Correct me if I am wrong, but didn't ROL have a license to RISC OS 4 and to its further development. "

Yes they have a license to update RISC OS 4, but there are limitations to that license about which no one who knows seems willing to divulge. ROL licensed RISC OS 4 off Acorn/Element-14. e-14 subsequently sold that "head license" to Pace - who had it for a few years - Pace then sold it to Castle. So in effect Castle have the *main* RISC OS license (which grants them access to updated apps/utilities (like later versions of BBC BASIC) and tools like the C/C++ which ROL *don't* have).

Andrew Banks wrote>"Has Castle now licensed ROL for RISC OS 6 ?"

No why would Castle who own the main RISC OS license need to license it off a sublicensee ROL. Besides as RO6 is really a 32bit RO4 and as Castle *already have* a 32bit RISC OS (RO 5) there's no need for them to do so.

Andrew Banks wrote> "The Guildford show this weekend should be very interesting"

The word "humdinger" springs to mind.

AMS on 19/10/06 1:41PM [ Reply | Permalink | Report ]

To clarify my last post "So in effect Castle have the main RISC OS license (which grants them access to the *OS* source and updated apps/utilities (like later versions of BBC BASIC) and tools like the C/C++ which ROL *don't* have).

SimonC wrote>"Castle don't seem to show the slightest bit of interest in anything non-Iyonix"

Er, yes - because in large part they sell the things ;-)

additionally could you imagine the row that would start if they tried to release RO5 for machines that ROL consider theirs... I'd also add your statement is untrue - the C/C++ tools are available to all (not just Iyonix users) and have been updated frequently. Also Castle make documentation publically available as well as !System updates and CallASwi module for download by 26 bit users.

CKH2>"Judging from the moderation scores, there appear to be lots (tens) of disgruntled Iyonix users out there who, admittedly, have received nothing yet from ROL."

As apparently have lots of non-Iyonix users who are Select subscribers who *also* haven't received anything recently from ROL. Hopefully that will change. I am one of those Iyonix users who frankly *doesn't care* whether Select comes to Iyonix - if it happens fine, if it doesn't I am no worse off really - you'll find overtime increasing numbers of us - and we'll be spending on RO5.12, updated C tools (another thing on my shopping list) and updates of the likes of TechWriter/Artworks - and ROL won't get a penny of that income - but then that's their choice isn't it ?

AMS on 19/10/06 1:53PM [ Reply | Permalink | Report ]

And also remember that the RISC OS market is not just desktops. Brittish Rail timetable screens on platforms used to have Acorn RISC OS in the bottom corner.

I have a bacteria incubator here that runs on WinCE. :-)

You get OS's in the strangest places. Cheers bob

nijinsky on 19/10/06 4:08PM [ Reply | Permalink | Report ]

Just for the record.. and I'm sure I'll be flamed for this but it needs saying:

Castle own RISC OS, not just a license to RISC OS, and this includes the head licence under which ROL operate.

The relevant contracts of sale between Castle and the previous owners, and contracts of sale various prior to that are, unfortunately, subject to NDA, so cannot be shown here. I have, obviously, seen and scrutinised them, having been closely involved in Castle's purchase, so speak, I believe, with authority.

ROL will undoubtedly say different.

All of which will just re-expose the argument. but.. that is life!

jb on 19/10/06 5:58PM [ Reply | Permalink | Report ]

jb>If your response is to my comment I am sorry if I misphrased it. I accept fully that Castle own RISC OS, and I was not attempting to imply otherwise.

AMS on 19/10/06 6:36PM [ Reply | Permalink | Report ]

jb:

So how do we persuade you and ROL to put aside your differences given where we are?

markee174 on 19/10/06 6:50PM [ Reply | Permalink | Report ]

Absolutely! Thats the question that everyone would like to hear answered please.

fwibbler on 19/10/06 6:56PM [ Reply | Permalink | Report ]

there's always going to be disagreement over this license, not least due to some of the ambiguous wording in it. However, that's rather beside the point. The problem is, and remains, Paul M. No matter what you say about Castle or any other RISC OS company, he, as druck says, lives in an alternate universe (having been present at some informal negotiations myself). Had anyone else been at the helm - let's say for the sake of argument and example, David Holden, I have no doubt at all that something, anything, could have been worked out - even if there were personality or business direction disagreements between any two parties. Such is the nature of business. Anyone who follows what I've said will know that at times, I've vocally disagreed with decisions that Castle have made - but in meaningful and constructive ways. Also true of other RO companies.

At the same time, I think that talk of "forked OSes" and the like continues to miss the point - such as that there are already multiple versions of the OS. As I said elsewhere, we must congratulate AD6 for taking charge of what they need for their machine, but ROL have alienated users and developers over and over. That could change, but I suspect it will not.

mrchocky on 19/10/06 7:04PM [ Reply | Permalink | Report ]

mrchocky: Nobody's missing the point about the forked OS except you it seems. There used to be several "static" versions that weren't being developed further with only ONE moving target to aim for. Now there are all those same static versions and TWO moving targets, each going in its own direction with no appearent cooperation. That's the fork people are talking about and I have a feeling you're intentionally just splitting hairs (no pun intended).

I do agree with you about ROL alienating users and developers. For the life of me I can't understand why people insist on failing to learn from history, even very recent events. ROL is steadily building a similar reputation as Microdigital, all talk, no delivery. Even the same arrogant "you are all wrong, we are right, although you happen to have proof otherwise but we'll still deny its existance" speech repeated over and over again.

BTW - I've gotten my hands on an old operating system that I'm going to update with all the latest and greatest features I can't list right now but it will run on old Ataris but will still be more up to date and advanced than any other modern operating system and give better performance as well. You can subscribe via regular mail, it only costs €150 a year and you might receive an update every now and then. If you're interested please visit www.payfornothingsucker.com.

Gulli on 19/10/06 9:51PM [ Reply | Permalink | Report ]

AMS:

"No why would Castle who own the main RISC OS license need to license it off a sublicensee ROL."

Because big parts of RISC OS 6 (everything that differenciates v6 from v3.8) belong to RISC OS Ltd?

"Besides as RO6 is really a 32bit RO4 and as Castle already have a 32bit RISC OS (RO 5) there's no need for them to do so."

Well, RO6 really is a 32-bit version of something like RO4.5, while RO5 really is more like a 32-bit version of RO4.

JGZimmerle on 19/10/06 10:11PM [ Reply | Permalink | Report ]

BTW, could the software-engineers who bug-fixed and modularised RISC OS 4 (and thus already know where the bugs were and how to fix them) simply do the exact same things again (in much less time, of course) on the shared-source version, legally?

JGZimmerle on 19/10/06 10:21PM [ Reply | Permalink | Report ]

Gulli: ah yes. As usual, users are insisting that developers don't know what they're talking about, despite their considerable experience with OS development. Thanks. I refer to you MW's missives on this very subject in several instances, where he agrees precisely with what I just said.

JGZ: Except that _both_ RO4 and RO5 contain a number of identical changes beyond 3.8. This is not surprising, unless like PM, we are trying to rewrite RO history.

mrchocky on 19/10/06 11:06PM [ Reply | Permalink | Report ]

RO5 was a marketing gamble which failed - deal with it. Castle purchasing RISCOS AFTER ROLtd had obtained a licence from Pace years earlier was a ploy to corner the market, which failed - deal with it. NDA's allow each party to hide behind the truth - deal with it and don't blindy assume you are being told the truth!

RO5 was released with new 32-bit hardware , the Iyonix, totally independant of ROLtd and RO4 because Castle thought they could corner the market. That it is a total no brainer which they will never admit to since that assessment failed.

RO4 (and now RO6) follows the same functionality as all previous Acorn OS releases in that it works on recent legacy Hardware. RO5 ONLY works on Iyonix and thus breaks that tradition. DUH. I can't believe how painfully in denial the Iyonix community is. Wake up. This truth has absolutely nothing to do with Mr Middleton's character irrespective of whether he is Mr Normal or as some have intimated Mr. Off-the-wall.

Clearly RO5 is the OS that has gone off at a tangent. However, a couple of years ago I DISTINCTLY REMEMBER Iyonix folks stating how ROLLtd would be dead within a couple of years and Iyonix/RO5 would dominate..........WRONG. It's wrong because the assessment of how the RISCOS market behaves is NOT like the PC market.

Now stop btiching, accept that mistakes have been made and be positive because absolutely f.... all will happen unless positive attitudes prevail. Recent developments have been very positive so why knock them or do some folks here wish to see RISCOS die a death ?

mripley on 19/10/06 11:15PM [ Reply | Permalink | Report ]

malcolm ripley: interesting assertions

1: The route taken by Castle was dictated by ROL's stance at the time that they were not prepared to 32bit RO, together with previous experience working WITH ROL (kinetic) which indicated a likely timescale totally unsuited to bringing a product to market. (Iyonix was approx 8 months from conception to launch!)

2: Castle made a business decision to BUY RISC OS when it was offered for sale, at a not insignificant price.

3: In deference to the licence that ROL do have (from Castle, the current RO Owners), Castle did not encroach on the ROL market for 26 bit machine OS, which is why RO5 is only on the IYONIX. STD had the opportunity to approach Castle for a 32bit version of RO, but for their own good reasons chose to go the way they did.

4: Cornering market: Competition in hardware is generally healthy. Please do not assume your (I presume) motives are Castle's

JOhn

p.s rose tinted spectacles are not generally compatible with investment.

jb on 19/10/06 11:34PM [ Reply | Permalink | Report ]

mripley: who's bitching now then? I'm not sure how you can say that "RO5 failed". The rest is rant, I wouldn't suggest anyone waste any time replying.

mrchocky on 20/10/06 01:55AM [ Reply | Permalink | Report ]

I hope that ROL are able to deliver RO6 on time. They seem to have a good product there - and a nice website in the works to boot, with useful developer info - so best of luck to them. It seems quite overpriced, though...

I think the comments in this thread sends a clear message to ROL - there is a substantial number of unhappy former, present and pontential customers out there and the company must keep their promises if they aren't all to disappear.

"RO5 was a marketing gamble which failed - deal with it." The early announcement of RO6, IMO, was yet another ROL PR failure. In a market fraught with vapourware, this announcement was never going to have much traction, particularly since the company has a reputation for promising and not delivering. If they had kept quiet and waited until they had a product to sell, I'm sure there wouldn't have been as much frustration from users. I can't help thinking that ROL are still miles behind schedule and that this was intended to keep users quiet - I hope I'm wrong.

The question is how many RPC/A7k users are there still out there, and do they outnumber Iyonix users?

hzn: PM seems to expect that customers, particular non-Iyonix users, will treat Select just like a charitable foundation - donate some money 'for the good of RISC OS'. Until RO6 is open source and has a transparent development process, I doubt many will fall for that - especially when the SSI gets going.

jb: Thanks for your informative comments, I'm sure it has cleared up things for quite a few of us. If only PM would do the same, I would like to hear his side of the story...

timephoenix on 20/10/06 03:53AM [ Reply | Permalink | Report ]

I wonder, if RISC OS 6 will support Cumana SCSI-2 Podules ? Because atm i cannot even make them normally working on RISC OS 4.39 Select 3 issue 4, except with ShareFS on my second RiscPC - Hopefully RISC OS 6 gets it working with the final release of it.

datawave on 20/10/06 03:59AM [ Reply | Permalink | Report ]

In reply to Drobe: Thanks for the update with the Select feature list link. I thought that there might be intersting info and for the heck of it tried ADFS which has a full 9 changes for S4. Well, after seeing entries like "Updated to build with ObjAsm and a modern makefile." as well as the odd small absolutely internal one. And for the odd data or resources module the change is "Updated header to claim 32bit safety." which does make sure there are lots of changes in number. What I would like to know is whom that list is intended for...

hzn on 20/10/06 05:52AM [ Reply | Permalink | Report ]

In reply to jb:

Thanks for your information posted here.

hzn on 20/10/06 05:55AM [ Reply | Permalink | Report ]

HZN: You appear to have missed this : [link] That's the "headlines" list of things of note for end users. ADFS has but one change, there. It's still a big list though.

imj on 20/10/06 08:43AM [ Reply | Permalink | Report ]

jb:

I actually agree with what you say but the appearance of 2 rival teams going their own way with different versions of RISCOS is not healthy and damages both of your businesses and confidence in the market.

The SE show gives you and ROL a marvellous opportunity to show some kind of unity of purpose and encourage us all to continue to invest and develop but you need to show some kind of interaction.

markee174 on 20/10/06 08:49AM [ Reply | Permalink | Report ]

Josh: "The question is how many RPC/A7k users are there still out there, and do they outnumber Iyonix users?"

All serious estimates suggest that a large proportion of Iyonix users also have older machines and that users with no 32-bit machines far exceed those with them. There are a fair proportion of users of pre-RiscPCs and, in my estimation, those at least equal the number of Iyonix users. That's ignoring those with 'a machine in the loft'. Iyonix users are more likely to use the Internet and (some of us) are more voluble in user groups making on-line estimates heavily weighted towards the Iyonix owners. If you are looking for people more likely and more able to spend money then look to the (comparatively) small numbers with Iyonix pcs. If you are looking for the greater untapped potential then look elsewhere where the numbers are much larger - but you'll need to work harder at selling the product.

jc on 20/10/06 10:01AM [ Reply | Permalink | Report ]

jc: "You'll need to work harder at selling the product" referring to older machines. I'd not bother putting too much effort into it at all. The very fact that they are still using 10 year old or more machines (if they are still using them at all) suggests they aren't keen on the idea of spending money in what's left of RISC OS. There might be more of them, but I doubt they are much of a market. Anyway, that's just my opinion, without any facts to back it up.

SimonC on 20/10/06 10:49AM [ Reply | Permalink | Report ]

In reply to John Cartmell:

ROL have to work even harder to sell a product that currently does not exist. I think many users will be keeping their cash fimly in their pocket until something is actually delivered. I myself a fully paid up Select subscriber, will not be resubscribing this year, unless they deliver on RO6 and Select 4.

sa110 on 20/10/06 11:04AM [ Reply | Permalink | Report ]

SimonC: You forget, that VA-RPC is programmatically a Risc PC, so if you're making a version for that, you get a version for the RPC/A7000 at the same time. Paul M at ROUGOL earlier in the year said that over 3000 copies of VA-RPC had been sold, that's probably a large part of their market now.

flibble on 20/10/06 11:13AM [ Reply | Permalink | Report ]

In reply to fibble:

And if you assume at least half of them upgrade to Select 4/RO6 that's a cool £148 500 (£99 * 1500). Not to be sniffed at. Should fund ROL for another year or two!

sa110 on 20/10/06 11:20AM [ Reply | Permalink | Report ]

In reply to SA110:

Don't confuse people who don't buy new hardware with people who don't buy new software. I'm a RPC user and have spent plenty of money on software and supporting the RISC OS market over the years. I also bet I'm not the only one.

Col1 on 20/10/06 11:24AM [ Reply | Permalink | Report ]

In reply to Col1:

I don't believe I have mentioned anything about RPC users. I simply took the 3000 VA users as an example and assumed that if half of them upgrade, it is still a considerable income for ROL.

sa110 on 20/10/06 11:49AM [ Reply | Permalink | Report ]

In reply to SA110:

My apologies- my comment should have been addressed to SimonC not yourself.

Col1 on 20/10/06 1:30PM [ Reply | Permalink | Report ]

Forgot about the VA users (depsite being one myself, as well as having an Iyonix). Anyway, it raises a point about ROL's "working with them" stance. No-one is working with them as regards to RiscPC / VA as far as I know, so why treat that any differently than the Iyonix? The obvious answer is because it's less work, but it still contradicts their position.

SimonC on 20/10/06 1:30PM [ Reply | Permalink | Report ]

In reply to imj: Thanks for that hint ... it might be a good idea to change the layout so that the long detail list is accessible through a link instead of being the most obvious thing on that page.

hzn on 20/10/06 2:26PM [ Reply | Permalink | Report ]

mrchocky: "ah yes. As usual, users are insisting that developers don't know what they're talking about, despite their considerable experience with OS development."

I have absolutely no idea how you managed to come to that incredibly creative conclusion reading what I wrote. Nowhere did I claim that developers don't know what they're talking about or even implied it, I was simply trying to clarify to you that what most people here are referring to when the talk about the OS fork, that being the ROL path vs. the Castle path and that you are talking about a different thing although related and that you were being your usual condecending know-it-all about it. Everyone knows that there are several old versions of RISC OS around that developers have to take into consideration, nobody has ever tried to deny that but that is not the issue being talked about. For once, try to let people speak for themselves, even when it doesn't fit with what you want to argue about.

Gulli on 20/10/06 3:42PM [ Reply | Permalink | Report ]

SimonC: "Anyway, it raises a point about ROL's working with them stance. No-one is working with them as regards to RiscPC / VA as far as I know, so why treat that any differently than the Iyonix? "

Since both Paul Middleton and Aaron of VA are travelling today I think I ought to correct your erroneous assumption. You couldn't be more wrong. There is a great deal of interaction between ROL and VA and between the VA programmers and the ROL programmers and this will continue in the future.

apdl on 20/10/06 4:52PM [ Reply | Permalink | Report ]

David: thanks for confirming to others that RO companies can and do, and regularly, communicate and cooperate with each other. The real situation isn't really like the perception that's prevalent on usenet and elsewhere.

Gulli: it's very simple, and you've demonstrated it again. You've dismissed my point out of hard without much reasoning at all. Sadly, something that's seen over and over again when it comes to developers presenting facts and opinions on things which they have considerable experience with. If users were to just give the smallest benefit of the doubt about such things, then we might not have much of the present situation. Read what ever else into what I say, and make unconstructive personal comments if you want, but the facts remain - the "OS fork" is much less of an issue that many believe or try of make out - for reasons, see many previous comments by MW.

"For once, try to let people speak for themselves, even when it doesn't fit with what you want to argue about"

Sorry, I have no idea what you're on about. How about sticking to stuff that's demonstrable rather than repeating things that "everyone knows" and making personal commentary. Here's one:

There are very very few applications (assuming they're 32-bit, and not for using hardware on a specific machine) that do not work on all RISC OS machines and RISC OS versions back to at least RISC OS 4, and often back to 3.5. And when differences _do_ arise, it's precisely the developers who have to deal with it. So, let's try not to be patronizing to those developers about what is and isn't an issue. As for "split" OS development - there are plenty of good real world examples of that - both Windows in the past and certainly Linux presently. I don't see those doing poorly.

mrchocky on 21/10/06 03:42AM [ Reply | Permalink | Report ]

mrchocky: What you say is absolutely correct, of course. The "OS fork" is pretty much irrelevant to developers of normal desktop preograms (and if you don't think I know what I'm talking about remember that APDL owns and is actively developing more commercial software titles than anyone else). Almost all the API's used by programmers are the same right back to RO 3.5 (and even beyond).

The only two things I can think of which are likely to cause difficulties are USB and the C libraries, and neither of these are actually OS problems.

I was writing programs for DOS when there was a *much* bigger OS fork betweed MS DOS and DR DOS. This affected hundreds of thousands of programmers and hundreds of millions of users. It didn't matter - as long as programs were properly written "by the book" they worked on either. In fact, there is absolutely no doubt that the competition lead both companies to improve their products. Perhaps it was not a probem simply because everyone knew that their software had to work on both OS's (and about half a million different sets of hardware as well).

apdl on 21/10/06 05:18AM [ Reply | Permalink | Report ]

In reply to mrchoky & apdl:

True, for the programmers the "OS fork" is not really an issue since the API is pretty much the same on all - assuming they are indeed written by the book. PM@ROL does try to make it an issue though by requesting programmers to make use of the new features of Select which then would result in apps running on Select only. But most programmers prefer to have their app run on as many systems as possible to they stick to the common API part which is the only sensible approach (until Select is there to have for all RISC OS computers).

But I think that for RISC OS as such the "OS fork" is a pity since it does result in duplicate developments thus using up more resources. And it certainly keeps the discussions about which RISC OS is better and the "OS fork" as such alive...

And as for Select and IYONIX pc: I don't want a full RISC OS Select - soft loadable Select features would be fine and probably much less work for ROL than supplying a full OS with all the hardware drivers and making sure that when the nVidia cards change that their OS runs on the newest IYONIX models...

hzn on 21/10/06 08:25AM [ Reply | Permalink | Report ]

Hzn: It's very easy to use Select features in an application without making an app Select only, and in fact no different to what we've had to do ever since RISC OS 3.1 really. The best example of this is the ImageFileRender utilities. Say you want to render an image selected by the user in your application (DTP, Web browser, whatever). Prior to Select you had to support file formats yourself. Instead of using OS_SpriteOp, or the JPEG calls, or the Draw calls, render the image using ImageFileRender IF the module is present. If not, fall back to the default behaviour. All of a sudden, your app can now support any type of image file thrown at it on Select. On RISC OS 5, the user can then have whatever the OS supports (or whatever your custom code supports). This is a couple of extra lines of code, so no big overhead. The benefits are clearly obvious by looking at Paint. A similar case applies to using semi-transparent images in documents etc - use it if it's there, disable it if the OS doesn't support it.

A good example of that is Steve Fryatt's Locate application. It works on all machines, but works as a filer plug-in on Select where it can. RISC OS 3 users similarly can use Configure plug-ins intended for RISC OS 4 - just not within Configure!

It's a misrepresentation to call it a 'fork' issue: that would suggest it's a new problem, and creates code which works on one OS but not another. Apps should work on all OSes, but there's no reason for them to not use features where available to them.

md0u80c9 on 21/10/06 1:45PM [ Reply | Permalink | Report ]

md0u80c9:

"A good example of that is Steve Fryatt's Locate application. It works on all machines, but works as a filer plug-in on Select where it can. RISC OS 3 users similarly can use Configure plug-ins intended for RISC OS 4 - just not within Configure! "

The irony, of course, is that when I first released a FilerAction Plugin version of Locate, I got questioned at length by ROL (or by somone claiming to act as documentation guru for ROL) about where I had acquired the bootlegged copies of the FilerAction Plugin documentation from. The documents I had used were actually supplied by ROL on the most recent Select CD, as I was a subscriber at the time, and were hardly dodgy (I had previously been under the impression that they were for use by developers, but it seems that I could have been wrong on that).

The attitude they took over the incident discouraged me somewhat from doing anything more with Select-specific code; subsequent events only reinforced that view.

stevef on 21/10/06 3:19PM [ Reply | Permalink | Report ]

In reply to stevef: With that attitude from ROL you describe and the information hiding of ROL (until recently when suddenly some programmers documentation was made publically avaibable) I would have refused to make use of any Select feature in my apps too... and re-thought about a select subscription as well.

hzn on 26/10/06 1:57PM [ Reply | Permalink | Report ]

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

Create a new account
Forgot your password?

Search this website

This week's poll

Recent developments have left me feeling...

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

Discuss this | Archives