Castle opens RISC OS future wishlist - RISC OS News, Software and Information

Castle opens RISC OS future wishlist

Published: 18th May 2004, 16:06:17GMT Source: drobe.co.uk
By Chris Williams

Page 1 of 1

CTL compares notes with users

Castle have published a meek road map of their view on the future of RISC OS development, which is about time too, given that they own it.

The RISC OS 5 developers have pledged to keep the Norcroft C/C++ compiler toolset up to date, but they say updating the RISC OS core is "commercially rewarding work" and "has the highest priority". Such development has, according to Castle, led to further support for ARM processor cores.

Future desktop enhancements, under the codename Merlin, will include: filer enhancements, extended graphics capabilities and network services, USB performance and connectivity, audio input, window manager enhancements, ROM application (Paint, Edit, etc.) enhancements and last but not least, bug fixes. These updates will be gradually released to end users, and Castle may introduce a pay structure for some features.

Castle have also opened a features wishlist, billed as a "6 week period of open consultation with the RISC OS community to assist Castle's management in allocating development resources through its internal Green Paper."

The Merlin website adds: "At the end of the 6 week consultation, Castle will begin processing the Wishlist database and compare it to their 'Green Paper'. Once processed, the Wishlist database entries will be made available along with Castle's comments and preliminary projections of future development plans."

The parallels between today's announcement and the early days of RISC OS Select are uncanny: RISCOS Ltd. also talked of C/C++ compiler updates, also opened a wishlist, also published a road map of desktop enhancements and also launched a scheme of periodic updates. Castle have also all of a sudden begun, uncharacteristically, pre-announcing activities and products: firstly, beta USB2 support is demonstrated at the Wakefield show, and then talk of un-named projects involving 50,000 units. As the divide between RISC OS 5 and 4 ever widens, Castle are either feeling the heat from the combined popularity of VirtualRPC and Select, or they have much to talk about.

Links
Project Merlin website
Press release

This article has been linked to, or is available in the following formats:

[Printable] [Digg this] [Blog search]

simo (+1.5) 18/5/04 4:54PM	How about a SINGLE version of RISC OS instead of the forking around 3.8? It's such a duplication of effort with two ADFSs, three, or four USB stacks..... It's funny how Castle seem to be the one in the corner though now as I'd have thought between Select for RiscPC, Omega, A75 and VA they must be selling more 4.x than 5.0x If Castle want to continue to develop RISC OS rather than just bug-fixing the Iyonix, then they need to buy ROL out, jees it's only a one-man-band!
harmsy (+1.5) 18/5/04 4:57PM	Sounds exciting. I agree with simo that it's odd to have two (or more) seriously different paths being taken.
Herm (+1.5) 18/5/04 5:00PM	Firstly- I applaud this initiative from the great people who brought us the Iyonix and, recently, the great looking Panther. Asking your target market what it wants is a good step forward. In the RISC OS field progress need not only to be made, but be /seen/ to be made. I feel that it is about time Castle learnt from the example of Select. Hopefuly this 'consultation' can lead to some of the features of Select appearing on 'Merlin' though duplication of effort within the platform does seem crazy. Maybe liaison with ROL can be one of our suggestions on the wish list (I know this requires a two way dialogue). That said I think that the community should get behind this initiative and provide constructive suggestions. Focusing on the negative can only be harmful for the platform as a whole.
adamr 18/5/04 5:10PM	I vote for them buying out ROL too! Ah, if only it was as simple as writing a comment on a Drobe forum Adam
j5m1th (+1.5) 18/5/04 5:14PM	Sounds like a good idea, and sounds a lot more than Castle "just bug-fixing the Iyonix". If they are actively seeking users opinions, then it sounds like they are really developing the OS. It is odd to have the OS going in different directions, and different companies developing each side. Of course the people to make the comments about the split to are Castle, and they are now asking for consultation with their users. Perhaps a chance to raise these concerns.
jonix 18/5/04 5:25PM	Marvellous! Let's provide a whole load of enhancements that have already been written. This argument is going to keep going round and round and round.
rod 18/5/04 5:27PM	Please please get talking with RISCOS Ltd and Microdigital. Let's have some dialogue going! We just need some putting together of RISC OS 4 and 5 that will suit all the hardware that we have in development. But then again perhaps this is just a consultation about the future of RISC OS 5, not RISC OS itself??!! Aargh!!! If Castle own RISC OS then take control of all of it dammit! I'm so tired of the messing around...
blahsnr 18/5/04 5:54PM	RO4.xx = emulators RO5 = new ARM based hardware Seems a simple split to me ROL won't have to write a HAL or 32 bit compatibility, Castle won't have to provide Windowsy GUI 'enhancements' I feel that we actually need is software new applications (Cino is a good example) to increase the number of things that we can do using RISC OS. Maybe changes to the OS will enable/enhance this process.
AMS 18/5/04 6:44PM	Good news then, nice to see Castle being proactive about engaging their customers (and potential customers). If I were them I'd advise against buying out ROL for a number of reasons: (1). What ROL have can't readily be moved to RO5 - so why pay out a wodge of cash for it and then pay again to have it adapted to RO5 (2). If Castle had to continue to support existing RO products under VA it would be a waste of resources and if they didn't it would probably Pss off people who bought into that "hybrid"* computer nonsense. (3). How pssed off would Omega users be if Castle didn't* support Omega as ROL had ? (4). All the above sounds expensive and that money and effort would be better spent on the hardware platform and the OS. The one unified hardware and software platform is the only sensible way for Castle to go (IMHO) that means Iyonix (or its successor) and RO5.
arenaman (-1.5) 18/5/04 6:51PM	The biggest mistake one could make here is to start putting "talk to RISCOS Ltd" on the wishlist. That is not an operating system enhancement. Here is a chance to put forth sensible and constructive suggestions as to what to add into the OS. It would be childish in the extreme to clog up the system with political and commercial comments. Castle are taking RISC OS forward. This further commitment to the desktop market is brilliant news. It's time to except that the future of RISC OS lies with this company. It has the money and engineers that RISCOS Ltd doesn't and to boot it has infinitely superior management and strategies, as has been demonstrated time and again. I can't see why anyone holds any sort of emotional or ideological attachment to RISCOS Ltd, who have, with the exception of releasing RISC OS 4, been mostly useless. RISC OS's glory days were when Acorn was developing the OS and the hardware together, which is precisely what you have now with Castle. With RISCOS Ltd, we ended up with some 56Mhz machines and a RiscPC with faster RAM (Kinetic). With Castle in control, you now have the option of buying a 600MHz XScale machine. It's a no brainer!! *In reply to blahsnr:* Agreed. Software development is key. But with Castle talking of further support of ARM cores and this open dedication to the desktop market, it seems a pretty stable commercial environment in which to release software.
dgs 18/5/04 7:06PM	*In reply to AMS:* "What ROL have can't readily be adapted to RISC OS 5" Oddly enough, Jack used almost exactly those words when he gave his presentation at ROUGOL last night. Castle and RISCOS Ltd have talked about ways and means of co-operation, but seem to have reached the mutually acceptable conclusion that neither will gain much by trying to re-transplant one company's development plans over the top of the other. Surely the userbase is better off accepting that reality, and pushing forward further development of RISC OS functionality, which is what Castle are offering - rather than diving into the politics of which company armchair critics believe should own which company. As blahsnr quite rightly said, ROL are developing RISC OS 4 and are likely to make most of their sales through PCs with emulators (plus the A75 that isn't really aimed at the ordinary user), and Castle are developing RISC OS 5 that is aimed at the end user who wants to use real RISC hardware. It's a neat split, and it all makes sense. What may start to upset people (perhaps has already?) is Castle increasingly stepping into Acorn's shoes. Selling tens of thousands of systems with RISC OS on real ARM hardware means Castle end up with vastly more resources and influence than anyone else in the RISC OS market, and the small players may quite understandably feel threatened. dgs
sebgate20 18/5/04 7:24PM	Does this ever mean RISCOS Ltd and Castle will co-operate?
adamr 18/5/04 7:35PM	arenaman said "RISC OS's glory days were when Acorn was developing the OS and the hardware together" Glory days my pants! "It was like Romeo and Juliet, except it ended in tragedy..."
simo 18/5/04 7:36PM	Well OK, if it doesn't make sense to try to port existing code between RO4 and 5, then at least make any NEW features version-independent - like one ToolBox, CLib etc. Having two ADFS's being re-developed for the same reasons at the same time is crazy. Just make one that will work on all machine types - if RPC, VA and Omega can use completely different hardware and all use RO4, then surely the Iyonix's much-heralded HAL can cope with using the same ADFS?!
AMS 18/5/04 7:47PM	*In reply to Simo:* I am not sure that the two ADFS developments are aimed at solving the same problem. Castle's ADFS will extend support for UDMA to CD/DVD devices (UDMA already works for the HDD). UDMA doesn't exist on ROL's implementation (it's not present on the VA implementation as the access is ultimately done through windows and Omega (to my knowledge) uses it's own IDEFS). If Castle had to take ROL's ADFS it would be a step backwards - CTL's ADFS already does UDMA - it would have to give that up to use ROL's ADFS wouldn't it ? And the HAL works under RO5 it wouldn't help port CTL's ADFS to the RPC either (quite apart from which the RPC doesn't do UDMA
AMS (-1.5) 18/5/04 7:48PM	Sorry 'bout the last sentence there - I'll try English the next time
nunfetishist 18/5/04 8:01PM	Well, UDMA is used when you use VA, of course, because Windows's IDE drivers do it. That, along with modern standards of block cacheing, is the reason VA blows everything else away in disc speed terms.
jms 18/5/04 8:04PM	*In reply to dgs:* "Surely the userbase is better off accepting that reality, and pushing forward further development of RISC OS functionality, which is what Castle are offering" That's what people said when ROL came along. Better support this company (ROL) which is pushing forward development of RISC OS in the wake of Acorn's demise. I bought Acorn, I bought Psion and now look where those companies are. I have this horrible fear that if I buy an Iyonix that (a) the £x00s I've spent on Select are well and truly wasted and (b) Castle could end up making STBs or embedded systems with no interest in the desktop market in a couple of years. I don't want to see my investment in RISC OS (the OS, not ROL) wasted. I don't want to buy a computer from a company that may in a few years have moved on to better, more profitable things.
AMS 18/5/04 8:11PM	nunfetishist> "UDMA is used when you use VA", yes I know, but re-reading what I wrote you'll see that I meant that the ADFS implementation provided by ROL is the bog standard one - it doesn't know how to do UDMA and that it is actually Windows (through VA) that does the clever stuff. Moving ROL's ADFS onto RO5 would be a nonsense as it is less advanced (relatively) than the one in RO5 already (never mind any further improvements CTL will make to it).
arenaman 18/5/04 8:12PM	*In reply to adamr:* A3000, A3010, A5000, A7000, RiscPC to mention a few of the machines that Acorn produced, along with necessary changes to RISC OS. Acorn machines were at one point far ahead of rivals in terms of speed and capability. Compare to what happened (or mor accurately, didn't happen) whilst RISCOS Ltd 'developed' the OS, and clearly the Acorn days were the glory days. Just because Acorn stopped making desktop computers and then disappeared doesn't mean the time before this was bad, does it?
arenaman 18/5/04 8:19PM	*In reply to jms:* Considering Castle are asking what desktop users want in the OS and are clearly commited to the desktop market, I don't see what you are worrying about. Especially since if you bought a Pee Cee for example, the same risk applies. Considering Castle are the ones developing new hardware and the OS to work on such hardware (such as IyonixPC), Castle own RISC OS and RISCOS Ltd clearly has no intention in looking to the future in this fashion (unless someone else pays the cost) then in my mind buying an IyonixPC is a better investment in the OS.
dgs 18/5/04 8:22PM	*In reply to jms:* "Castle could end up making STBs or embedded systems with no interest in the desktop market" You need to listen more carefully to Castle presentations. That's not Castle's aim, and they certainly don't have the naive approach that Acorn had regarding big external customers hopefully descending on them with big chequebooks. Yes Castle do want to take advantage of investment from larger companies, but the reason they're putting huge efforts into the enthusiast market is that thousands of enthusiasts spending a thousand or two pounds each is still a very worthwhile amount of money - that's not going to change fast. Do you know of any companies who got rich making set top boxes? I don't. Watch this space for a more detailed examination of some of what Castle have been doing and saying recently - it may meet some of your concerns. Incidentally, I've been a Select subscriber all along, I've bought an Iyonix and am considering buying a second Iyonix. I don't consider my Select subscription wasted, I could always keep a RiscPC and put RISC OS 4.39 on it for fun. dgs
mavhc 18/5/04 8:28PM	As long as the APIs are the same so you won't need two versions of programs. But of course there needs to be the same bug fixes and features so that programmers can use them over the whole platform. If Select features were standard for Iyonix, VPC and Omega users and an option for RiscPC users more programs would use them. We're basically going to have 3 groups of users, ROS4, ROS4.5 and ROS5, so everyone will still have to write for ROS4 users.
adrianl 18/5/04 8:29PM	"That, along with modern standards of block cacheing, is the reason VA blows everything else away in disc speed terms." Well, that and the fact that that you're probably referring to the RISCOSmark benchmark which doesn't test the disc access speed at all under VA because it just repeatedly loads a 1MB file which has been cached by Windows and is thus little more than a memory copying test. If that's what your RO app happens to be doing then fine, but it's unlikely to be representative of most apps.
jms 18/5/04 8:40PM	*In reply to dgs:* "I don't consider my Select subscription wasted" If you no longer use a RPC (or only rarely) and Iyonix is the way forward, then what benefit has years of spending money on Select provided? None! As a 'normal' user I find the situation of two different versions of RO seeming to pull further apart utterly depressing. Anyway, I look forward to reading the upcoming Castle article.
arenaman 18/5/04 9:02PM	*In reply to jms:* That's like saying what's the point of buying a new car now when in a few years you will replace it. It doesn't mean the car was a waste of money. You buy what you need or want. If you used it until it was time to move on, then it wasn't a waste of money. Apply same logic to RISC OS.
thegman 18/5/04 9:16PM	I think I agree with jms, an easy analogy would be that if you were a Mac user, you used Mac OS X 10.0, then 10.1, 10.2, and then 10.3 with your Power Mac G4, but when you upgraded to a G5 you had to go back to 10.0, 10.3 was not compatible. Having now used an Iyonix, I can see what the Select fans are talking about, whereas before, I quote frankly did not see what the fuss was about. Select has some stuff (like alpha-blended PNGs) which allow you to achieve effects which are just not possible on an Iyonix. Not until (if) Vantage appears on it, anyway. That's not to slight the Iyonix in any way, it's cool, it would just be cooler with Select.
jms 18/5/04 9:31PM	*In reply to arenaman:* I'm not quite sure I follow the car analogy ... but if I use it, I feel it's like buying a new car that doesn't have the aircon, electric windows and GPS system of my current car, but it'll have a bigger engine and uprated suspension. I might then have to wait another couple of years for this model to get aircon, etc. Personally, I prefer a car with aircon to one with a big engine.
dgs 18/5/04 9:34PM	*In reply to thegman:* Comparisons with other operating systems are not always appropriate. Other operating systems have a tendency to get slower with every iteration. (It's one of these sad facts, even despite developer efforts to avoid it). But most RISC OS developments, including RISC OS 5, tend to make things faster with new releases, not slower. Long may that continue. If the G5 ran at twice the clock speed of the G4, with the cache, disk access and so forth to back it up, and similar features, then I doubt anyone would be using a G4 Mac even if they were the only ones to have Mac OS X 10.3 or whatever. (Does this comparison look silly now? It's your comparison). dgs
dgs 18/5/04 9:46PM	*In reply to jms:* It's more the case that the Iyonix does have the air-con, and electric (front) windows, and cruise control and remote central locking. These are standard features (USB, networking, DHCP?) that some other cars are only just getting. The Iyonix also has the large engine, obviously. The Iyonix arguably doesn't have the shiny R18 alloy wheels, the racing stripes, and the ten-CD autochanger as standard. Maybe Castle's PANTHER pc is the racing stripes one. If your current car has a GPS navigation system, and you use it all the time, then you definitely don't want to do without something similar in your next car. Is there anything that is similarly vital in your Select software, that isn't provided with the Iyonix? i.e., can you be specific? Are we fed up with car analogies yet? dgs
Use the forum for more comments on this article