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

"Vast majority" of ROS 4 now 32bit

By Chris Williams. Published: 19th Apr 2005, 21:41:32 | Permalink | Printable

And I'll do things my way

RISC OS Ltd. motifRISCOS Ltd. has claimed that it is close to producing a 32bit RISC OS 4, independent of rival Castle. The news comes after earlier rumours that plans to merge RISC OS 4 and 5 have stalled. ROL have also announced that they intend to allow people to subscribe to a 32bit Select in early May, and will begin development of Select for the Iyonix range of computers once they get 100 punters signed up.

"All new RISC OS 4 code has of course been developed to be 26/32 bit neutral with the aim of producing a 32 bit version of Select as soon as market conditions allowed. This has been expanded throughout 2004 so that the vast majority of RISC OS 4 source code - i.e not just modules which have been enhanced via Select - is now 32 bit safe," RISCOS Ltd. said in an email to Select users this evening. The company says it will produce 26bit versions of Select for legacy machines, such as RiscPCs, and 32bit versions of Select for newer computers.

The announcement added: "This has been a major undertaking and it should be pointed out that this work has been done completely independently from the 32 bit version of RISC OS that had been developed by Pace that was used by Castle for RISC OS 5."

Around this time last year, RISCOS Ltd. called on Castle to cooperate in the efforts to produce Select for the Iyonix. The Iyonix was launched under much fanfare in 2002, as its 32bit RISC OS was seen at the time as a breakthrough for the market.

Links

Select 4 outlined

Previous: Cineroma demonstrated in public
Next: RISC OS Memory Protection

Discussion

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

so we'll see RISC OS Select 32 for Iyonix sometime around .... 2030 perhaps :-( , the latest (and fastest) machine the RISC OS market has to offer (the Iyonix) is still at the bottom of Paul Middletons to do list, or maybe im wrong, although i'm sure i'll be corrected :-)

 is a RISC OS Userleeshep on 19/4/05 9:52PM
[ Reply | Permalink | Report ]

With a 32 bit RO 4 now in existence, hopefully it will not take them long to make it work on the Iyonix. As to it been at the bottom of their to do list. Personally I'm not surprised. If the majority of their business has been 26 bit, then it makes sense for them to concentrate their efforts on that market.

 is a RISC OS Usersa110 on 19/4/05 10:02PM
[ Reply | Permalink | Report ]

Nice to see things moving on.

 is a RISC OS UserRevin Kevin on 19/4/05 10:22PM
[ Reply | Permalink | Report ]

will this open the door for xscale Omega??????

 is a RISC OS Usermicrobits on 19/4/05 10:23PM
[ Reply | Permalink | Report ]

leeshep: I don't think you need to maintain such a pesimistic outlook. I can only see this news as great, spectacular and not-before-time. :)

I'm going to divulge into pure speculation and predict that Castle will have a Iyonix Select within the next month or so. Why? The Iyonix is fast becoming the machine of choice for most RISC OS users and the market for 26bit versions of RISC OS must have reached saturation point. ROLtd need more customers not just subscription renewls. From the wobbly fence I'm sitting on I think that the only way they can do this at the moment is to develop a 32bit RO with Iyonix users in mind. Why Iyonix users? Most of ROLtds Select additions have been cosmetic, or user-interface enhancements (well, the best ones have). And where are the most desktop users of RISC OS? On the Iyonix.

As usual I'll probably be proved completely wrong and they've probably ported RO to x86. :)

 is a RISC OS UserSnig on 19/4/05 10:29PM
[ Reply | Permalink | Report ]

So, how does one interpret this?

Are the two RISC OS camps going to continue to march in different directions?

 is a RISC OS UserGulli on 19/4/05 11:41PM
[ Reply | Permalink | Report ]

"...it should be pointed out that this work has been done completely independently from the 32 bit version of RISC OS that had been developed by Pace that was used by Castle for RISC OS 5."

I wonder wether I should applaud RODL for this costly and demanding work or pity them. Why wasn't RO4 & 5 merged one way or another (seeing as they both emerged from Acorn's RO3.8)? Perhaps the forthcoming Select 4 would then be available for the Iyonix as well as 26 bit machines...

 is a RISC OS UserhEgelia on 19/4/05 11:55PM
[ Reply | Permalink | Report ]

"Why wasn't RO4 & 5 merged one way or another"

As has been explained many times before, the two branches of RISC OS are now rather different from each other and it would not simply be a case of automatically merging the differences, it would be a "costly and demanding" work in itself.

 is a RISC OS Userjms on 20/4/05 9:14AM
[ Reply | Permalink | Report ]

Does this mean that ROL will have to develop their own HAL and USB stack for use with the Iyonix? This would seem to be a fair bit of work, and it makes no sense at all, since perfectly good ones exist in RO5. Deleveloping the USB side of RO5 has not been without its pain, with some devices stopping working at various points in its evolution. Have we Iyonix users got to go through all that again with Select for Iyonix? (or whatever it is going to be called) Or will sombody knock Castle and ROLs heads together and get them to see sense?

 is a RISC OS Usermrtd on 20/4/05 9:22AM
[ Reply | Permalink | Report ]

The Cineroma article mention a rumour about a new APDL for the Iyonixs, and here ROS 4 is near 32 Bit complete.

Maybe because the RO4 and the RO5 user markets are so independent that the developments are still that way with a vision of merging in the developing time line near future?

 is a RISC OS UserSawadee on 20/4/05 9:23AM
[ Reply | Permalink | Report ]

I can definitely confirm that there will not be a new APDL.

 is a RISC OS Userapdl on 20/4/05 9:40AM
[ Reply | Permalink | Report ]

Sawadee APDL=ADFS?

 is a RISC OS Userblahsnr on 20/4/05 9:42AM
[ Reply | Permalink | Report ]

"Does this mean that ROL will have to develop their own HAL"

The email did mention development of "Video Hardware abstraction" but made no mention of a HAL or USB stack.

 is a RISC OS Userjms on 20/4/05 10:31AM
[ Reply | Permalink | Report ]

HAL may not be necessary

What is necessary for Iyonix is a driver for the video and USB

 is a RISC OS UserROHC on 20/4/05 11:03AM
[ Reply | Permalink | Report ]

"The email did mention development of "Video Hardware abstraction" but made no mention of a HAL or USB stack."

Not really surprising with regard to the USB stack since 32-bit Select's development is probably payed in the most part by the company behind the A9 and which can offer its own USB stack.

 is a RISC OS Userandretim on 20/4/05 11:43AM
[ Reply | Permalink | Report ]

jms: Sure, it has been 'explained' before, but not to my and other users' satisfaction! Why exactly were plans announced to merge both branches and then apparently withdrawn? I understand the two branches have evolved in dissimilar paths, but why in the first place? Why the diversity? In some way I expect RODL and Castle to explain to their customers why this had to happen and why users had to be, and are, asked for more money to replicate hard work.

So, now we may be expecting a 26 bit Select for machines with both a 26 bit and 32 bit processor mode and a 32 bit version for CPU's with only the 32 bit mode, such as the XScale. Why not a 32 bit version for all users relying on 32 bit software (both Acorn legacy machines/Omega and Iyonix/A9home) and a 26 bit version for those still working on 26 bit only software (like me)? Aemulor developer Neil Spellings has even suggested a version of Aemulor could be made for Select 32 users on 26/32 bit CPU's (ARM6/7/SA) needing to run 26 bit software occassionally and it would use the CPU's native 26 bit mode, thus preventing significant speed / performance losses. Please correct me if/where I'm wrong here.

 is a RISC OS UserhEgelia on 20/4/05 12:05PM
[ Reply | Permalink | Report ]

Sawadee I suspect that a 32 bit Select/Adjust for a computer not made by Castle is far more likely than Select/Adjust for the Iyonix (or indeed any Castle produced computer). Acorn and Apple were/are successful when they controlled both the hardware and OS (in the case of Acorn when they had control over CPU manufacture). Castle now having arrived at a similar position are unlikely to relinquish this freedom.

My feeling is that if the A9Home becomes a fact, you will have to buy that instead of an Iyonix to get Adjust features. The other options is to buy/install the 3rd party utilities for your Iyonix that get you to somewhere near the Adjust feature set. I am not sure what you can do to get RO5 only features on a Select/Adjust machine.

Even if there were compelling technical reasons to merge both source trees, I think the politics of the situation make it very unlikely that the two camps with be able to co-operate.

Being very happy with RO5 I can't see a good reason to subscribe to Select 32. I'd prefer to spend my money on new products such as a Real or Windows Media Player for RISC OS or a video editing suite rather than paying for the rather pointless duplication of OS development.

 is a RISC OS Userblahsnr on 20/4/05 12:39PM
[ Reply | Permalink | Report ]

After using Select and Adjust for a couple of years, the step back in features with my Iyonix is really limiting.

I was begining to regret the purchase, but the announcement is very good news.

 is a RISC OS Userjess on 20/4/05 1:15PM
[ Reply | Permalink | Report ]

well when applications start saying "Needs RISC OS Select 32" and/or "Needs RISC OS 5 or greater" then we know that they might as well be different platforms.

Windows is going through this right now with 64 Bit. I have come across a couple of apps / games that need 64 bit version of windows to run / use extra features, and it isnt out fully

but hey why put all the money into one deveolpment pot for the platform, might as well make a third one, to split the money 3 ways

 is a RISC OS Userem2ac on 20/4/05 1:44PM
[ Reply | Permalink | Report ]

jess Just goes to show how different users have completely different usage patterns and thus very different perceptions of the same products :)

 is a RISC OS Userblahsnr on 20/4/05 1:45PM
[ Reply | Permalink | Report ]

hEgelia: "Why not a 32 bit version for all users relying on 32 bit software"

I think the ROL argument goes along the line that (a) running a 32 bit OS on a RPC would break pretty much all network / podule cards, (b) users would be required to buy (at extra cost) Aemulor or equivalent to run any existing 26 bit apps and (c) ROL would then have another version of the OS to support.

From what Adrian Lees has posted to the csa.*, my understanding is that he feels that Aemulor would run at a reasonable speed on a RPC and may even speed some software up (by optimising the code as it goes).

 is a RISC OS Userjms on 20/4/05 2:12PM
[ Reply | Permalink | Report ]

I think Jess' comment sums up my reticence to purchasing an Iyonix. Now once the latest version of the OS i.e. select/RO4 is available for the Iyonix I'd purchase one. I suspect I might not be alone with that plan. One proviso, if Omega with Xscale card is available (unlikely) then it might get purchased instead.

 is a RISC OS Usermripley on 20/4/05 2:40PM
[ Reply | Permalink | Report ]

Why would a 32 bit RISC OS cause any trouble on a RiscPC? I thought that the 32 bit issue meant that the code was compatible with a fully 32 bit processor, not that it required one, so a StrongARM could still run in 26 bit mode, running a 32 bit compatible OS, as well as other 26 bit code, without needing Aemulor.

 is a RISC OS UserSimonC on 20/4/05 3:43PM
[ Reply | Permalink | Report ]

jms: (a) That's a very good point, but perhaps at least some podule drivers could be re-written for 32 bit compatibility? At least the network card drivers, since they are already included in the OS? Otherwise Aemulor could help?

(b) ROLD are planning to release 26 bit Select anyway. I'm just saying give users the choice: Do I need 26 bits backwards compatibility or not? If not, they can go 32 bit in full gear.

(c) From what they seem to be planning now, they already would be supporting at least 2 versions of Select and a version of Adjust. IMHO it would be a waste of money for users to buy Select 26, then buy Iyonix or A9 and pay again for Select 32, probably with a small discount. OTOH, ROLD would earn more money this way and inject that in much welcome OS development.

Perhaps Drobe can do a poll inquiring on users' need for 26 bit compatibility? I just wonder how many people still need to run 26 bit only apps (of which I'm one ;)).

If Aemulor could even speed up progs in certain cases...

 is a RISC OS UserhEgelia on 20/4/05 4:06PM
[ Reply | Permalink | Report ]

The speed increases mentioned come from Aemulor (in theory) being able to optimise applications for a given processor type. Given a large number of applications that Aemulor is targetted at are written or compiled for ARM2/3/610 processors, they arent StrongARM or XScale aware, so are not optimised for those platforms.

We have always said that if ROL were to product a single Select32 for all platforms we would do a cut-down Aemulor to handle the 26-bit apps. As a few people of mentioned, RiscPCs can run both 32-and 26-bit code, but if the OS is running purely in 32-bit, then 'something else' would be needed to handle the transition of the two worlds and ensure they coexist happily.

However, ROL have always maintained they have no plans for a pure 32-bit Select for legacy machines for many of the reasons stated in this thread - app compatibility, podules etc.

My main concern is that if a Select32 is ever produced for the Iyonix, how much of our other software (Geminus, Cino etc) would break because they uses features only provided by RISC OS 5.

Regards,

Neil. www.spellings.net

 is a RISC OS Userspellinn on 20/4/05 5:18PM
[ Reply | Permalink | Report ]

ah there we go....some real use of said thing

I think that ROL, or what ever they are called these days, produced it not for the iyonix, that would be much better, then HELP Castle develop RISC OS 5...notice that word....HELP

Meh! whose going to listen to a user?

 is a RISC OS Userem2ac on 20/4/05 5:36PM
[ Reply | Permalink | Report ]

The problem is Castle's agenda for RO5 is vastly different from ROL with Select. Castle are pushing RISC OS (rightly so) into high volume PDA/STB markets where there is real money to be made. Theres no reason for ROL to help them for that market, as all ROLs developments are (rightly so) targetted at the desktop market.

ROL's problem is that they are relying on a dwindling number of RISC OS desktop users for their main revenue stream. Depending on how sucessfull other OEM licencees are (A6 etc) in widening the market for RISC OS machines could well determine whether ROL sinks or swims.

Cheers,

/Neil/ www.spellings.net

 is a RISC OS Userspellinn on 20/4/05 5:45PM
[ Reply | Permalink | Report ]

to spellinn: not sure I'd agree that the 26-bit revenue stream /is/ dwindling - what about all those emulator sales? In fact a twin-track strategy appears feasible: 26-bit for legacy hardware and VRPC users, 32-bit for A9 and - hopefully - Iyonix users. I would have thought that the latter are by now too big a constituency to ignore. It would be incomprehensible to withhold the most highly developed version of the OS from the fastest hardware. Cheers George

 is a RISC OS Userbucksboy on 20/4/05 7:09PM
[ Reply | Permalink | Report ]

I don't know whose fault it is but I'm just deeply depressed by this. Our OS is crying out for modern features and facilities and what we get is *two* separate 32-bit conversions (not a trivial job since we have previously been told it is impractical).

And the duplication will go on. Why can't we see a division of labour that delivers one core 32-bit OS and a whole bundle of nice enhancements? Nice enhancements - there's a whole load of excellent ideas in Merlin for starters (no-one need get their egos on the line, they're user's ideas, not Castle's or ROLtd's). Commercial logic even suggests that, by each doing something different, they each get more customers.

I think we've heard all the reasons why not and they don't stack up. These undoubtedly clever people need to take a step back, see that the market needs progress, not duplication, and MAKE IT HAPPEN!

Sorry for the rant. It won't happen again. I just hate waste and want to continue using RISC OS in the long term.

 is a RISC OS UserTonyStill on 20/4/05 9:23PM
[ Reply | Permalink | Report ]

Can't see what the fuss is. I always liked RISC OS due to the combination of the user interface and the top quality applications. Also, it being mostly ROM-based and fast and reliable was a huge plus, too. When I use any computer, it's not to ogle at some operating system, it's to use software running on it, so if the OS doesn't provide a graduated pindrop or multi-user accounts, it doesn't bother me. It's a pretty simple choice anyway - if you want greater speed and capabilities, choose IyonixPC and RISC OS 5. If you want bells and whistles, stick with your RiscPC and RISC OS 4. If 32bit Select appears for RISC OS 5, and you feel it is worthy of purchase, buy it.

 is a RISC OS Userarenaman on 20/4/05 11:45PM
[ Reply | Permalink | Report ]

hEgelia: "Do I need 26 bits backwards compatibility or not? If not, they can go 32 bit in full gear."

You might also ask "what benefit will 32 bit RO provide on a RPC?" Given the additional costs associated with upgrading software, buying Aemulor and losing various podules.

"IMHO it would be a waste of money for users to buy Select 26, then buy Iyonix or A9 and pay again for Select 32"

I suspect that, given the different underlying hardware, that a Select 32 that runs on a RPC won't run on an Iyonix (unless all drivers for all hardware are to be supplied with Select 32).

 is a RISC OS Userjms on 21/04/05 09:50AM
[ Reply | Permalink | Report ]

Either way, My RISC PC is at least getting RISC OS 4 :D might go the whole hog and get select, its the last upgrade it will EVER get (well aparet from emory, i think the 256 its got isnt enough, its running slow! :@P

 is a RISC OS Userem2ac on 21/04/05 10:41AM
[ Reply | Permalink | Report ]

em2ac: And how will putting in more memory make it faster, even if you could?

 is a RISC OS UserSimonC on 21/04/05 10:53AM
[ Reply | Permalink | Report ]

32 bit risc os on a risc pc should alow bit wimpslots. It would also allow developers with only a risc pc to be sure their apps are 32 bit clean.

I think the idea would be a neutral version of RO (for RPCs) that by default comes up in 26 bit, but can be configured (CMOS) to start in 32 bit or to ask the user before loading.

 is a RISC OS Userjess on 21/04/05 10:59AM
[ Reply | Permalink | Report ]

oops BIG wimpslots

 is a RISC OS Userjess on 21/04/05 11:00AM
[ Reply | Permalink | Report ]

In reply to APDL:

Sorry about that, I did mean ADFS.

 is a RISC OS UserSawadee on 21/04/05 11:09AM
[ Reply | Permalink | Report ]

"You might also ask "what benefit will 32 bit RO provide on a RPC?" Given the additional costs associated with upgrading software, buying Aemulor and losing various podules."

There definitely are benefits to it, one already pointed out by Jess. As to your other comments, I think I already answered them when I said 'give users the choice'. If you'd lose too much or are too dependent on 26 bit RO, then stick with it and buy 26 bit Select if you like. The only thing I'm saying is that there probably are users who'd like 32 bit RO on their RPC or Omega and won't necessarily lose compatibility when ditching 26 bit.

I think a forthcoming Select 32 for the Iyonix would be quite different from the current Select 26 anyway, because of the different hardware involved. I also think ROLtd should not have much trouble in compiling a 32 bit Select for the RPC (and possibly Omega) seeing as the article already stated that a lot of RO4 code, and Select enhancements, are 32 bit clean. I can't be sure of that, ofcourse.

 is a RISC OS UserhEgelia on 21/04/05 12:25AM
[ Reply | Permalink | Report ]

I thought that the whole point of having a HAL was that ROS 5 did not access the hardware directly, so that changes to hardware could be easily done without causing operating problems. All that a new computer would require was its own HAL to work. Have I misunderstood the whole theory?

At the last Guildford Show it was mentioned in the Joint Castle/ROD Talk that Castle had produced a HAL for the RPC, during development work on the Iyonix.

Surely 32 bit Select for the Iyonix would need to support the HAL system, otherwise it would be a backwards step.

 is a RISC OS Userajb on 21/04/05 1:13PM
[ Reply | Permalink | Report ]

HALs rarely do more than memory mapping, interrupt routing, and such. Doing that bit is the easy part. You've then got to write drivers for everything, which is somewhat more involved.

 is a RISC OS Useranon/194.83.236.250 on 21/04/05 3:16PM
[ Reply | Permalink | Report ]

Well, Castle must have decided that making use of a HAL for the Iyonix was the best system.

Wouldn't driver writing be required whatever method was used ?

 is a RISC OS Userajb on 21/04/05 5:13PM
[ Reply | Permalink | Report ]

I have been reading all the comments up untill now and they all seem to make sense in a way, but personnally I have made a disenchanting observation, which I would like to share.

It is almost a year now since I have upgraded to Select 4.39 and have not renewed my subscription after that. I use Iyonix as my main machine and, after completely crashing on my Kinetic RPC, Select now runs on an A7000+.

The few times I have to use the A7000+ I have become more and more irritated by the (mostly cosmetic) differences which now are getting in the way rather than making things more easy, because I am so very much used to RO5 and that is why, on these occasions, that I mostly revert to 4.02 now.

Whether or not Select and RO5 are developing in the same or separate directions, to me that is of no importance at all. Hm, makes me wonder, am I the only one?

 is a RISC OS UserIke on 21/04/05 8:31PM
[ Reply | Permalink | Report ]

Adam: I'm confused about your assumption that Select for the Iyonix would be completely different from any other Select. Apart from the links to the hardware they would be identical being derived from exactly the same source. That surely is what it's all about. Whilst the interface layer will be different the user shouldn't be aware of differences - except that more software will run on the RiscPC than will run on an Iyonix or A9. And at a different speed!

 is a RISC OS Userjc on 22/04/05 01:46AM
[ Reply | Permalink | Report ]

Ike I went as far as Select 4.37 and found the changes particularly to the networking part of Configure most irritating whilst trying to set up a home network with a new router (and subsequent changes to the network). I have downgraded one Select RiscPC to 4.02 (with Iyonix desktop icons), the other one if not sold soon will go the same way.

I was very pleased when I got the Iyonix that RO5 looked and felt more like RO3+ than RO4.02/Select . Keep the interface pure and allow third parties to write the utilities they want and users to install what they want.

 is a RISC OS Userblahsnr on 22/04/05 08:39AM
[ Reply | Permalink | Report ]

The advantage of the HAL is that it isolates the I/O funtions from the OS kernel, allowing the same kernel to run unchanged on machines with differing I/O systems. It provides a fixed interface to the kernel for I/O devices, even though the I/O devices themselves may have different interfaces. So yes, you need a driver for each type of device, and that still has to be written, but this can be done without affecting to OS kernel if you have a HAL. IMO a Select for Iyonix that didn't use the HAL would be a backwards step, even though the Select kernel will probably need major changes to interface with Castle's HAL. Incorporating the HAL will probably cost more in the short term, but will make things easier in the future. I am also concerned that Select for Iyonix might lack some features that exist within RO5. If so, like Ike, I might prefer to stick with RO5 since it does most of what I need. The most important missing features under RO are mostly to do with applications and are being addressed. So for Iyonix users who have never used Select, this is not a burning issue. RO5 is perfectly usable, and is much better than RO3. I can understand that a lot of work has also gone into RO4 that it would be a pity to lose, much of it invisible to users, but RO5 required work to produce as well. I for one would like to know what Select for Iyonix offers that justifies the cost of subscription when there are so many other things I could spend the money on that would be more useful to me, like supporting Peter N's Firefox project for example. My fear is that the reality of a even a merged Select/RO5 might be less than the sum of its parts. But ISTM that we are being offered an Iyonix version of Select, grown from RO4 without any input from RO5. This could lose many RO5 features that those used to them would miss, just as much as many people used to Select find RO5 deficient in some areas.

 is a RISC OS Usermrtd on 22/04/05 09:09AM
[ Reply | Permalink | Report ]

Well I now have RO4.39 and think its much improved over 4.02 which itself was much improved over 3.7 (i.e. RO5). The thought of going back makes me shudder. However some people don't like change even when its better.

 is a RISC OS Usermripley on 22/04/05 09:16AM
[ Reply | Permalink | Report ]

Simon Challands:

I was joking, you know....having a laugh...this would be the solution on PCs

 is a RISC OS Userem2ac on 22/04/05 09:41AM
[ Reply | Permalink | Report ]

Malcolm,

for the very last time: RO5 is not RO3.7. Not even remotely. Not from a technical point of view. Not from a user's point of view. If anything, it is very similar to an advanced version of RISC OS 4. Think of RO4.02 with Unicode support, DHCP, USB2.0, LanManFS with long filenames, large WimpSlots and FAT32 support.

In contrast, RO4.39 is not similar to an advanced RO5, because it has a few additional features, but also lacks a few (IMHO very important) features.

Steffen

 is a RISC OS Userhubersn on 22/04/05 09:47AM
[ Reply | Permalink | Report ]

mripley It is about usage patterns. I didn't find anything in Select for my usage pattern that helped me (and a number of things got in the way). You might find that strange but there you go.

To be honest the only two things in the GUI I really miss in RO3.11-3.7 compared RO4/Select 4.37 and RO5 are long filenames and the iconise functionality. The latter is the only one that really annoys me and there is a 3rd party utility for RO3+ that helps.

 is a RISC OS Userblahsnr on 22/04/05 10:19AM
[ Reply | Permalink | Report ]

em2ac: When posting from work I'm very unlikely to be in the mood to see even the most obvious jokes, and to just take things literally, so accept my apologies :-)

 is a RISC OS UserSimonC on 22/04/05 10:22AM
[ Reply | Permalink | Report ]

What features does RO4.39 lack compared to RO5, other than those required to support the Iyonix hardware? Unicode fonts is the only thing I can think of.

 is a RISC OS Userhelpful on 22/04/05 10:29AM
[ Reply | Permalink | Report ]

John: "I'm confused about your assumption that Select for the Iyonix would be completely different from any other Select. Apart from the links to the hardware they would be identical being derived from exactly the same source."

Well, 'completely' different is a bit strong, I say quite different because, like RO5, major parts of the OS had to be rewritten for the different architecture involved. Now, ROLtd has been hard at work ridding RO4 of its dependency on VIDC20 and IOMD while developing Select and even though I think they've come a long way in realising a full abstraction, they're not quite there yet. In this sense I think there probably are still some underlying parts which need work before Select would run comfortably on Iyonix hardware. From a users' point of view I guess there would be little difference from 26 bit Select as we know it, although I think it would be a safe assumption to say ROLtd would take the opportunity to offer some special advanced features tailored for the Iyonix. In the end, I could very well be mistaken and what I'm saying here is purely speculation on my part.

 is a RISC OS UserhEgelia on 22/04/05 10:35AM
[ Reply | Permalink | Report ]

hubersn ...not forgetting UDMA100 support for IDE discs, PCI graphics and USB cards and support for Gigabit networking.

helpful There was a thread a while ago that RO4.xx versions could not make full use of higher speed network cards due to limitations in the network stack. RISC OS development (excluding Select) has always occurred to support new hardware and that is when significant changes have occurred to the OS. Without support for new hardware RISC OS will only live running on emulators.

 is a RISC OS Userblahsnr on 22/04/05 11:00AM
[ Reply | Permalink | Report ]

As a daily user of the Adjust ROMs on a StrongARM and an Iyonix (Networked together), it's the Select enhancements to !Paint and !Draw that I have most enjoyed. I use !Paint and !Draw a lot, and I always hop back to the StrongARM to do so. But the Iyonix is faster, and when an app needs a bit more speed it's real nice to have the extra boost the Iyonix can provide. BASIC, in particular, is very much faster, more so than you'd expect, on the big I. Just a ramble on my usage patterns.... Martin Hansen.

 is a RISC OS Usermartin on 22/04/05 4:04PM
[ Reply | Permalink | Report ]

Whenever Select does come for Iyonix i will buy one and i am sure that is true for other users? Castle take note!

 is a RISC OS Userpmcd on 24/04/05 01:30AM
[ 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

  • Review: A9home v. Koolu
    Clash of the tiniest
     31 comments, latest by polas on 18/10/07 6:03PM. Published: 15 Oct 2007

  • Random article

  • ROS-powered school looks for careers tutor
    And an ICT technician
     Discuss this. Published: 24 Apr 2007

  • Useful links

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

    Top developers:
    RISCOS LtdRISC OS OpenMW SoftwareR-CompAdvantage SixVirtualAcorn

    Dealers:
    CJE MicrosAPDLCastlea4X-AmpleLiquid SiliconWebmonster

    Usergroups:
    WROCCRONENKACCIRUGSASAUGROUGOLRONWUGMUGWAUGGAGRISCOS.be

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

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


    © 1999-2009 The Drobe Team. Some rights reserved, click here for more information
    Powered by MiniDrobeCMS, based on J4U | Statistics
    "Drobe doesn't reflect the entirety of RISC OS as it declines to publish some news stories because 'they are not interesting' "
    Page generated in 0.3985 seconds.