Revin Kevin (+1.5)
15/11/04 10:36PM |
With the 32bit select demand just asking the select subscribers is a bit limiting as if someone just has an Iyonix then they are unlikey to be a select subscriber but would possible be interested in the 32bit select.
|
martin (+1.5)
15/11/04 11:24PM |
Well, as an Iyonix user with an Adjust ROMed RiscPC I'd buy it right away if it were available for Iyonix; and that without whatever Merlin ends up adding to the OS.
Personally, I'd prefer to hand over cash and get Adjust32 on a CD in exchange - I'm just old fashioned that way - but I can see why RISCOS Ltd prefer subscribers funding on-going developement. Adjust certainly shows that RISCOS Ltd can deliver on their promises.
Crumbs; can you imagine them turning round and saying there's insufficient interest ? I've no idea of the level of enthusiasm voiced for Merlin. Was that alone not warm enough to suggest that there was the demand from the Iyonix user base to make a Select32 (including Merlin) worthwhile ?
It must be one hell of a task to merge this forked OS - They are wise to tread carefully - Presumably the continuing success of Virtual RiscPC must make RISCOS Ltd wonder if their time would not be better invested developing that further.
Looks like 2005 will be another year of interesting twists and turns for users of our favourite opperating system. |
jmb (+1.5)
16/11/04 12:28AM |
I'm a little disturbed by this development.
Firstly, how can ROL expect to gauge the "real level of demand" for a product if they don't actually ask the relevant people?
Secondly, the impression I got from the SE Show was that the Merlin wishlist and relevant information was to be handed over to ROL. If ROL are now hoping that Castle will quietly drop Merlin and leave them to do desktop development, it suggests a lack of communication (which I find hard to believe as Paul M was in the room when Jack L stated that the details were to be handed over).
Thirdly, and to my mind, more importantly, if ROL turn round and decide there's insufficient demand for a 32bit Select (or whatever) it leaves the future of RISC OS on the desktop looking rather bleak (as the desktop OS developer would be showing a distinct lack of interest/resources in moving the platform forward). |
rmac (+1.5)
16/11/04 1:38AM |
I assume, perhaps incorrectly, that you need an Iyonix to utilise a 32bit select?
Could one assume that only some of the current Select users/subscribers have an Iyonix? If you do the survey seems seriously flawed by ignoring the largest pool of Iyonix owners i.e. the ones without Select.
The only people with a more complete list of Iyonix users, the most likely customers of 32bit select, is Castle. Why aren't Castle/ROL using their customer database to get an accurate picture of possible demand?
It almost seems like a survey designed for the purpose of establishing a low demand for 32bit Select - virtually guaranteed without surveying the owners of Iyonix machines. Will this then be used to justify NOT 32bitting Select? |
tamias (+2.5)
16/11/04 2:06AM |
In reply to rmac:
The StrongARM is a 32-bit processor with a legacy 26-bit mode which RISC OS 4 uses.
So you could theoretically run Select32 on a RiscPC. Most likely the 32-bit version of Select, once merged with RO5, would implement RO5's HAL (how else would it run on the Iyonix?) so if ROL decided to make a compatible layer for Select32 to run on IOMD/VIDC in the RiscPC then it would run.
But then you have to ask yourself, why would you want a 32-bit OS on the RiscPC, particularly if it provides no extra features over Select26? As on Iyonix, 26-bit software would break and have to be run under an emulator such as Aemulor. The RiscPC is simply too old and slow to make this feasible.
I agree about the rather limited scope of the survey though - surely the Merlin wishlist is enough for ROL to realise that there is demand, but most of their potential customers are still just potential customers so there's no use asking only existing customers what the potential customers would like! |
blahsnr (+2.5)
16/11/04 6:56AM |
I have an Iyonix but have not been a Select subscriber for a while (last installed version was 4.37)
The problem for me is not just whether or not Select 32 provides 'significant' improvements over RO5 but also the fact that I decided not to buy products again that require deposits paid or subscriptions in the hope that something useful is delivered the money you place in advance.
Until there is a finished product (like Adjust) then I can only ever be a potential customer.
I realise that people have found Select on a RiscPC to be a great improvement. For my useage patterns, Select did not provide any noticeable improvement over RO4 and a number of disadvantages. RO5 is pretty much what I want, fast stable and a clean simple desktop.
If aspects of Select/Adjust that make my Iyonix go noticeably faster were to be included without the multi-user logon (and the associated additional complexity in the Select !Boot) and GUI 'enhancements' then I might be interested in those improvements.
Ultimately the biggest problem for me is having to buy products from a company that I decided I wasn't going to buy from again. |
chriswhy (+1.5)
16/11/04 8:30AM |
I have just received a letter from RISCOS Ltd confirming my renewed Select subscription (dated 14th Oct) attached to and invoive and credit card slip (dated 17 September). So I know I am on the subscriber list. But I haven't received any emails about a 32bit Select. Mind you from the time it took to confirm my Select subscription, I'll probably have to wait until mid December.
Do I want a 32bit Select? Yes, I've got an Iyonix as well. What I would really like is a roadmap for RISC OS, show what is planned for the next 1-2 years in the same way other companies provide roadmaps for their OSs. That should indicate when Merlin and other features are in the development priorities and when releases are planned.
In the meantime I'm getting to likeLinux on Iyonix, having bought it at the SASAUG Show. maybe that's where the future lies. |
Sawadee (+1.5)
16/11/04 8:51AM |
ROL should know how many Iyonixs have been sold? Surely, wouldn't this info and Merlin Wishlist info would be shared between the "Now in Partnership" RISC OS companies?
In one view, ROL are only asking the current market for a gauge of demand on Select32, on the other hand it sounds a bit of a daft question pointing to an uncertain RISC OS future "if there is sufficient demand" ? To justify 32bitting Select by the current user market doesn't make sense, forward thinking to produce Select32 for the future RO market and it's growth irrespective of considering the effective production cost factor for now!
I strongly believe the future of the RISC OS market is our future growth, and not just us.
We are not a big number to gauge any financial forecast on really?
I wonder how costly it really is to produce Select32? |
mrtd (+1.5)
16/11/04 9:12AM |
I think that the problem for those of us who have never experienced Select is to know what we will be getting for our money. RISC OS 5 already does what I need it to do, so I need a good reason to buy Select 32. However, the ongoing development and support of RISC OS is important, and if there are worthwhile improvments (ie not just cosmetic or things that I don't need like multi-user login) I will almost certainly support it.
As an Iyonix owner who has never used Select, I would like ROD (as I suppose we should call them now) to send me a mail shot or even an email describing the existing Select features that will be included in Select32, plus their planned future developments and asking for support.
If there are significant improvements existing or planned, I, and I am sure many other Iyonix users like me would be happy to subscribe. I think ROD are missing an opportunity here. |
jlavallin (+1.5)
16/11/04 9:15AM |
Sounds like ROL are trying to wriggle out from producing Select32.
Perhaps I should just buy an Iyonix and quietly forget ROL exists
With the promise of XScale for the Omega there is another interest group to be added to Iyonix users.
|
nunfetishist
16/11/04 10:00AM |
Running a 32bit OS on a RiscPC may have a use: RiscPCs are somewhat cheaper than Iyonixes if all you want to do is test your software builds for 32bit compatibility. |
jess
16/11/04 10:35AM |
other benefits of a Select + RO5 on an RPC:
RO 5 allows huge memory slots,.
It would be the same OS as new machines would have.
If people (who otherwise would stick to 26 bit) migrate to a 32 bit OS, then it would encourage even more 32 bit updates, which must be good for the platform as whole. |
bucksboy (+1.5)
16/11/04 10:59AM |
If Select is not 32-bitted, it will only be usable in future under emulation (VARPC or a future Linux equivalent), which would result in the somewhat bizarre consequence that the main product of the company responsible for the platform's OS development wouldn't run on native hardware.
Now it may well be that an emulation-only strategy is practicable and defensible, even desirable, but if that /is/ ROD's business plan then I would hope they would make that clear. But their track record does not inspire confidence that such a pro-active approach will be adopted. Nonetheless, it is most important IMO that our main desktop OS developer has a clear strategy which they are prepared to defend and promote as required. Hopefully that will embrace modern native hardware in addition to any other solution.
George |
Spriteman
16/11/04 11:09AM |
CTL: OK, ROL, you can look after the desktop side of our OS and provide new versions for the Iyonix
In reply to ROL:
Sounds good to me.
...some time late...
In reply to ROL:
Anyone interested in 32bit Select? No? Didnt think so. Job done. Man, I'm good. |
CJE (+1.5)
16/11/04 12:21PM |
Select on the Iyonix is fundamental to ROD's (ROL's) plans for the future so I don't know where "ROL is trying to wriggle out of doing Select on the Iyonix comes from". They could have worded the posting better. As to why ask on the Select mailing list surely that is the obvious place to start (i.e. existing customers), I expect they will also ask on the Iyonix mailing list.
I do understand those people who are reluctant to use 'beta' software, and releases do need to be managed, but to some extent Castle have been doing that for the last two years and have not attracted criticism!
In view of the complaints (unjustified I believe) about Castle's recent email shot ROD may have been a bit wary! |
jess (+1.5)
16/11/04 1:00PM |
Hasn't select already been 32 bitted? A9 range.
The question is whether it will be ported to the Iyonix or it will be a new OS combining the best of both select and RO5. |
md0u80c9 (+1.5)
16/11/04 1:28PM |
In reply to Jess:
Indeed. The question is actually more relevant than it appears. Select32 exists - but not Select for Iyonix. The two are different concepts.
Remember that running on an Iyonix requires two components: a 32-bit OS and suitable hardware-dependent code to run upon the machine.
ROL without question have a 32-bit version of Select; this is already running upon the A9 series as has been announced by Advantage Six's press release. My understanding from that press release is that they are able to ship to their corporate customers machines running Select32 on an Arm 9 based machine now. Advantage Six would no doubt have been involved in developing software to control the hardware specifics of their machine.
However, just because it runs on one design of Arm 9 machine or runs on the RiscPC doesn't mean it runs on any ARM based system (which is why Select32 wouldn't run on my phone, if I'm being silly about it!), and it is further development work to then develop for Iyonix hardware.
That work would therefore need to be done to integrate Select 32 with the Iyonix hardware, and that needs to be funded by the potential profits.
Clearly ROL are selling Select to Iyonix users not Castle, so no doubt it would be ROL who would be footing that development bill. Therefore enough users would have to want it to be worthwhile. |
adamr (+1.5)
16/11/04 1:34PM |
In reply to md0u80c9:
OK, that all makes sense to me. But Castle have passed the batton of "desktop" development to ROD, right?
So, presumably, whatever the specifics of the question on the select list, the real question that they are asking is: "Do you want any further developement of a desktop version of RISC OS?"
Or have I missed some other option here? (Sorry if I'm being stupid!)
Adam |
xyzzy (+1.5)
16/11/04 2:56PM |
It makes sense to ask current Select subscribers whether they are interested in a Select32 as development of 32 would have to take a chunk of time/money away from the current Select. Having a ballpark figure for the number of current subscribers wanting to switch to Select32 or add it in to their current sub allows some idea of how much funding may be available before looking outside the current Select 'fold.'
People with experience of Select can then more realistically know which of those features/improvements they would find useful and expect to see in Select32 and so it might enable development to be pitched towards what a majority would want.
I'd hope that this is just the start of the process of getting the show on the road. The message was only sent at 2140 last night and seeks to gauge the "real level of demand" and "features expected of it".
They state that "much of what was proposed for the Merlin project was not of general interest" suggesting they have worked through the Merlin list and now seek sensible suggestions.
As we still have RiscPCs here, I'd be in the (b) add-in group. |
bucksboy
16/11/04 5:13PM |
In reply to Jess:
you're right of course: Select32 does exist - for the A9 family. The problem is, the impression that was given when peace broke out between ROL and Castle of a coherent strategy for the future whereby Castle would concentrate on the embedded market and ROD would do desktop development already appears to be fragmenting, and ROL/ROD to be reverting to type, i.e., 'give us a wad of cash and we'll do you a version of the OS'. Hence Select32 for the ARM9. But what about ARM10/11/X-Scale?
Castle's strategy, of developing an OS core-plus-HAL, at least had the merit of being adaptable to other versions of the ARM, as I understand it. ROD's apparent strategy will simply replace hardware dependency with software ditto.
If I'm talking through my hat, please say: it wouldn't be the first time.
George |
mrchocky
16/11/04 6:00PM |
No, Select32 doesn't exist - Embedded Select32 does. The distinction is a matter of effort, which this is presumably to fund.
However, many of the proposed Merlin improvements are really quite different to those in Select - both are important in different ways, but I'm not so sure about "choosing" between them.
|
druck
16/11/04 6:06PM |
From the answers Jack & Paul gave at the SE show presentation, Select on the Iyonix will initially consist of softloaded components, untilising the exsiting HAL and hardware dependant drivers of RO5 (so all existing Iyonix hardware will containue to work).
Wotk on merging the Select and RO5 source trees will eventually result in a single ROM image incorporating both the RO5 HAL and the Select enhancements. This will still be softloaded on the Iyonix, and a vanilia RO5 image will remain in flash as a fall back. Although eventually when mature it there could be an Iyonix Adjust which would be committed to flash. |
md0u80c9
16/11/04 6:14PM |
In reply to MrChocky:
agreed that this is strictly true; the definition between what constitutes 'embedded RISC OS' and what constitutes 'desktop RISC OS' is blurry; I'd maintain the issue is more one of hardware compatibility than missing components. I think we would have to wait until a review or larger description of the A9 came out before we could comment on how much of 'desktop RISC OS' is present in the A9 series.
In reply to bucksboy:
Where is your hat  . The issue of processor support is one, but probably not the biggest issue by any means. I would suspect running on an XScale is the least of the problems.
A HAL is exactly what you describe: a 'Hardware Abstraction Layer'. It allows you to carefully define with as little code as possible, differences in machine architecture. All OSes have hardware dependency - 'ROD's apparent strategy will simply replace hardware dependency with software ditto' therefore makes no sense. Linux, to quote the 'daddy' of interoperability has fantastic hardware support, but still requires work to function on a new system. It is, to all intents and purposes, hardware dependent.
What a 'HAL' /isn't/ is a magic wand allowing RISC OS to work on any ARM machine thrown at it. RISC OS 5 stands as much chance of working on my phone as 4.37 does in that respect. The lack of suitable hardware support is the reason why unmodified 5.0X won't work on a RiscPC.
I'm not in a position to comment on whether or how complete Select 32's hardware independence is, BUT a HAL doesn't mean it's easy to port to any machine you like; it just means you provide entry points for a potential hardware manufacturer to insert their code. You still need to write code for every processor or controller chip in the system, and that takes time and money.
Normally the person doing that coding is the machine supplier; Castle are in the position that they don't need to do this as they're shipping 5.08. Therefore, ROL needs to pay for that development. |
lproven
17/11/04 3:23AM |
Sod porting an existing version of RISC OS onto a machine which /already has one/.
Surely it would be more use to try to move it onto ARM-based hardware which doesn't currently run RISC OS? For instance, there are many ARM-based machines which run WinCE. These can run other OSs - there are several versions of Linux which run on them to some degree, e.g. Familiar, Debian and OpenZaurus. Ergo, there is freely-available information out there on how to access the h/w of these machines. If there's any commonality to that platform, a port to this would be more use. While RISC OS might not be much use on an iPaq, it would be very useful on subnotebook-class machines, such as the Compaq Aero. Alternatively, Simtec's ARM development boards spring to mind.
If something must be dropped to free up development effort, then I'd ask why maintain Norcroft C/C++? Might it be possible to move to GCC instead, thus getting the benefit of a lot of external development work? BeOS did this around the move from r4 to r5, and it is, for example, possible to compile a compatible, working Linux kernel using Intel's C compiler instead of GCC. |
blahsnr
17/11/04 6:58AM |
How will this work? Exactly the same way as with Select on 26 bit machines?
That is to say IyonixPC's will continue to ship with the current version of RO5.0x and then as a user you can choose to buy a replacement Select version?
Or will IyonixPC's in the future start shipping with the current (stable) version of Select-32? |
JWCR
17/11/04 9:59AM |
I really think that Select-32 as a seperate ongoing project is an irrelevence. The advances that ROD have produced in Select-26 should be made available for the Iyonix (or any new 32 bit machine that might become available), as a one off paid for upgrade. If RISC OS Developments is to be the guardian of desktop OS, it should be properly financed to develop and upgrade the unified OS for all users |
bucksboy
17/11/04 11:38AM |
mdOu80c9: thank you for pointing out the non-sequiters in my post. In effect, RO5 and Embedded Select32 are not as dissimilar in concept as I thought. But the implication of ROD's question still makes me uneasy: why spend time establishing the /real/ level of support for something you've committed to do anyway? And if you're possibly not going to do it (as a result of the response), what then?
George (plus hat) |
md0u80c9
17/11/04 12:46PM |
Then presumably they have other work on, and the question is simply which should take priority. If you were sat around with nothing to do, you wouldn't ask whether it's worth doing something . |
chriswhy
17/11/04 12:55PM |
I have had my Iyonix for nearly 2 years, during which time I have received over 40 upgrades from Castle (excluding betas). If I don't subscribe to Select32, will I stop getting updates to RISC OS 5? When?
I have suscribed to Select(26) since it started and received 3 releases and 1 set of ROMs which I had to buy on top of my subscriptions. OK, the Select releases contained a lot more change that 1 of Castle's upgrades, but you have to wait a long time between releases for bugs to be fixed. besides my RPC is now shut down most of the time so I am not sure I need Select26 any longer.
If very few of us subscribe to Select32, does that mean no more development of RISCOS 5? Does it mean no more development of 32bit RISC OS in any form?
Given that in the past ROD have said a lot of the changes are under the bonnet in Select26 and that prevented providing the visible bits (Paint, Daw, Edit etc) for RISC OS 5, just how easy is to provide partial upgrades vias softloaded modules before a softloaded Select32 for Iyonix is available?
I would relly like to see a roadmap from ROD for RISC OS, the 26 bit and 32 bit5 select streams and RISC OS 5, what comes together when, what new developments are expected and what releases are planned. I am not asking for an exact list of developments for each release with exact shipment dates, rather a prioritised list of planned developments and which quater/year they are penciled in for. Other companies do this, why can't ROD? |
| 1 comment(s) are below your moderation threshold. Login to view them. |
| Use the forum for more comments on this article |
RISCOS Ltd. contacted its 
