Revin Kevin
2/8/04 9:48PM |
They have no need to defend a good product if other people are to blinkered thats there problem. |
fylfot
2/8/04 9:52PM |
Thanks for this article. It's reminded me that we should be oh so grateful to STD for continuing to manufacture Windows PCs for the RISC OS community. Their talent for making Windows seem like a cool product deserves some respect. And they're just so community spirited. |
TonyStill
2/8/04 10:09PM |
The original quote on Drobe referred to the A5 as a "proof of concept/technology demonstrator". I, like others, I guess, assumed that the technology being being demonstrated included the compute platform. That appears to include the WINTEL bits.
Personally, I am attracted to an ARM-based platform because I like the overall RISC OS/ARM conbination. If the technology being demonstrated is the RO/tablet parts, with the final product being a battery-efficient ARM-powered machine, then that's much more attractive.
However, I stand by my comments on the attractiveness of an XScale-powered PDA; not a directly comparable product but at least as attractive.
Either way, I wish STD luck in moving RO forward into new niches. |
adamr
2/8/04 10:26PM |
In reply to tonystill:
I am attracted to an ARM-based platform because I like the overall RISC OS/ARM conbination.
Why, for Pete's sake? Have you got some strange attraction to running a desktop computer on a chip designed for a mobile phone? Apart from sheer bloodymindedness, of course? 
Adam |
thesnark
2/8/04 11:25PM |
In reply to adamr:
Err... you do know that the ARM was designed for a desktop computer, right? |
nunfetishist
2/8/04 11:28PM |
In reply to thesnark:
The original ARM was. ARM havn't designed a CPU in donkey's years for desktop use. 99% of ARMs designed and fabbed these days are designed for mobile phone use, RAID cards, or other embedded applications. Certianly not desktop machines.
I'm sure STD would be happy to build a true-ARM based tablet, or high-performance desktop machine, if somebody gave them a huge wodge of cash. But the market can't really support that much new hardware considering the cost of developing it. |
jess
3/8/04 12:41AM |
In reply to adam:
I want an arm based desktop, for similar reasons to using low energy light bulbs - efficiency. nun: STD don't need to build one, just to get RISC OS running on an already existing one. |
adamr
3/8/04 1:02AM |
In reply to jess:
As I understand it (and I reserve the right to be wrong!) Intel Centrino (?) processors use less power/MIP than "high performance" ARMs... |
jess
3/8/04 1:17AM |
Can't see how, arms are far far simpler, so any improved technology applied to them as well would restore their lead. I can't see intel missing out in the phone market and not applying advances in efficiency. |
simo
3/8/04 2:22AM |
What's wrong with using up-to-date hardware to run our out-of-date operating system, when there's no better native hardware alternative? |
NoMercy
3/8/04 3:38AM |
(you know I'm geting narked off at firefox, every time I accidentally hit my nav buttons on my mouse my commend disapears)
I'm not going to believe anything I hear about Pentium M/Centrino's Mips/Watt figures, since after 30 minutes of searching I find absolutely nothing apart from brags about how there 'high', but I don't really have any worries that the Pentium M will cut into ARM's market share in the hand-held devices front, but it might pinch Transmeta a bit, who currently sit on the sub-notebook area and look to take a portion of any Tablet PC's which come out (would be nice to get X-Scales in a few wouln't it)
I suspect however that ARM's are stuck on low-power devices forever and the only viable option to turn RISC OS into a powerful platform operating system would be to migrate to PPC , x86 wouln't present any comercial opertunities for other markets of the system board and IA64/etc are all rather expensive and dificult to implement. |
Sawadee
3/8/04 3:39AM |
I agree with "simo's" point, at least STD may provide for user's who need emulation to substitute what we don't have and cannot provide for now.
The flip side of the coin is the three wishes of "Merlin's" magical "Modernisation" project which looks good.
Then we shouldn't complain about the choices we have with the two sides of the RISC OS coin, as long as it isn't a double headed coin then we could end up with more choice than what Windoze offers?
Regards, Steve. |
epdm3be
3/8/04 5:29AM |
We (users) should have realised from the start that the way to take RISC OS forward was to blindly fund Windows XP machines without giving them indication as to whether the lot would support all the onboard features as/if provided by the underlying Windows XP operating system and make sure that their intended application runs secure and stable on top of a highly flawed OS in the first place.
Yep this IS the way forward. Standing inline next to emulators of defunct home- and gamecomputers running on top of an acclaimed insecure OS.
BTW. fylfot: STD doesn't "manufacture" Windows PC. They're no more than a box-shifter in this respect (just like all the others with their shuttle XPC's orso).
Oddly enough it's been told here several times that perhaps if they really want to use excisting technology; Why not use X-scale Pocket PC's, PALM's or those X-Scale linux PDA's orso. That would be a nifty "Technology demonstration" and a far more interesting "Proof of concept". Just MHO of course.
EPDM |
epdm3be
3/8/04 5:31AM |
On the other hand.
Why do we really care?
At least that same STD gave our aging RPC another 10 year lifespan with that multipod, isn't it?
So it can't be all bad. |
mikeg
3/8/04 7:45AM |
As the original target of Stuart's newsgroup reply, I need to wave my oar, don't I. I've made it fairly clear that I can't see why STD and RComp need to badge bundles of VA and clone PC as if they were some special box.
I don't object to the VA product at all. It's the bundle that's the problem.
We have banged on about the reliability of "our" kit for years, reduced Cost of Ownership justifying higher purchase price, etc. I know that sticking VA+Adjust into the mix ramps the base price by a bit. It does seem strange, though, to go for bottom-end machines to demonstrate the "proof of concept". Either STD will be selling them, in which case it matters to a degree in terms of margin, or they won't. If the A5 is a "proof of concept", surely a better base would have been more sensible?
Maybe it's the ability to badge the clone in question, although that begs the question of the transparency of the demonstration. Does one pretend there is nothing but RISC OS on the box, and not let the client touch it? Or is there a big "ta-Daa" moment, when the emulated nature of the beast is revealed?
Bottom line is I so don't like the smoke and mirrors bit. There's nothing wrong with the idea of demonstrating the concept of RISC OS on a stylus-entry device using emulation as the first step on the way to a native pen-based device. It's a wonderful idea. The point of the branding exercise escapes me, though. |
jess
3/8/04 8:14AM |
In reply to epdm3be:
Making RO work on an existing X-Scale PDA device would require quite a lot of work. Making VA run on a similar looking windows device does not. Therefore you show the VA system to potential buyers (outside the current market), saying the real thing will start quicker, be lighter and last longer, if enough are interested you pay for the RO port to the X-Scale PDA, if not you sell the VA unit to the existing market. |
mripley
3/8/04 8:29AM |
If its a proof of concept the concept is a "RISCOS Tablet". As STD says there is no point in spending buckets of customers money and their own to provide, on delivery day, a RISCOS Tablet which the customer looks at and says "What's this ? Where's the Start button !" etc etc. Instead the customer can see how a RISCOS tablet would work without anyone spending one red cent on hardware..... Now if the customer is happy with the "concept" then a full RISCOS/ARM version can be built.
As for the ARM is for phone comment, well honestly! Here's a simple question if AMD had the opportunity to provide an x86 chip for phones would it say "No thanks we make chips for desktops not lowly phones" or "Yes please (whilst the dollar signs flash before their eyes)" ? The market is changing rapidly from desktop based computing to hand held computing. That "phone" is slowly turning into an all encompassing PDA and multimedia device. The latest phones have games, MP3 players, PDA stuff and the internet. Everything in fact that the vast majority of people rushed out to buy a PC for a few years back. The x86 family is struggling to meet that demand due to its high power demands. Oh by the way for those obsessed with meaningless MIPS. It's not the MIPS that a phone company worries about its the total power requirement. Since it's the total power consumption that defines the battery life and if you have two chips with the same total power requirement you then choose the one with the higher processing capability, hence why the ARM chip is very popular. |
jess
3/8/04 8:36AM |
Another thing that it neatly demonstrates to new markets is the fact that any purchaser could run the same software on a PC via an emulator - just like Palm Pilots did (or is that do?) VA could be an importan part of any connectivity pack |
JWCR
3/8/04 10:12AM |
It seems quite simple to me. STD is saying:
1) Here is an ARM powered device running Windows
2) Here is the same device running RISC OS under emulation. Good, isn't it.
3) Did you know that groovy RISC OS doesn't need Windows to run on an ARM device.
4) If you give us your money, you can hae this ARM device with all the advantages of RISC OS
and none of the disadvantages of Windows.
Simple, a proof of concept, a way to wean people who are so used to asking, as mripley says "What's this ? Where's the Start button !" |
jess
3/8/04 10:19AM |
Start button - maybe it would be good if there were an App (NOT part of the OS EVER) that provided a start button in the bottom left, alt-tab, tree save, windows tool order and jump to front windows, to ease the transition. |
chrisj
3/8/04 10:44AM |
In reply to Jess:
There are numerous pd/freeware/other programs around that do provide alt-tab, app-launch from the icon-bar, and so on - either all-in-one or separately; take your choice.
In reply to NoMercy:
(miles further up)
If you want a really powerful RISC-OS machine, the way to go is support for multiple processors. There are already a number of cheap chips containing several (usually 4) ARM cores and suitable support; 4x600MHz would be "only" about 2GHz (maybe even a touch less) once you allowed for switching losses and the like - but the power consumption (and price!) would be much lower than for an X86, so you could always fit more than one.... |
quatermass
3/8/04 10:59AM |
What about this Transmeta Cursoe Chip?
It's suppose to be software programmed to behave like another processor. So why not a ARM chip?
http://www.transmeta.com/crusoe/
There are a number of laptops and Tablets using this processor.
http://www.transmeta.com/success/tablet.html
So I wonder if it would be possible to take a complete Tablet PC, reprogram it to be a ARM chip and then rewrite the kernel so it can communicate with its hardware?
|
nunfetishist
3/8/04 11:14AM |
In reply to quatermass:
The code morphing software that runs on Crusoes is just as complex as any other JIT, as that's what it is. You'd be much better off using a faster CPU.
What's all this about VRPC funding Windows? I imagine it's such a TINY number of sales, nobody at Microsoft have noticed. |
jess
3/8/04 11:18AM |
In reply to quatermass:
have you noticed what processor the A5 uses? If it is possible, then perhaps STD are one step ahead? |
jess
3/8/04 11:36AM |
nun: But do faster processors do this in "hardware"?The point quatermass was making that you could have what is effectively a StrongARM machine and create a version of RO for it. (presumably you would include FP too). That arrangement would be of interest to me. |
NoMercy
3/8/04 12:15PM |
Ichrisj:
I did a little bit of looking at the MPCore by ARM, and in raw MIPS it comes to around the same as a 1GHz x86 chip, though I failed to find any implementations of it... would be an interesting study to see how much preformance at what price could be achieved by building a system with 4, 8 or even more processors  |
nunfetishist
3/8/04 12:21PM |
In reply to jess:
No. They do it in software. Just like the Crusoe does. |
AMS
3/8/04 1:51PM |
In fact ARM are starting to realise that their "turf" may be invaded by others (cut down Mips and PPC processors and even (gasp) x86 derivatives) so they're gradually having to up their specs. Allowing for larger/faster caches and so on (I don't have the URL to hand but ARM do discuss larger fast caches - which would improve performance) and there are multi-core ARMs (or even single core faster ones such as the 80331 mentioned here on drobe).
The capabilities of ARM will improve (as ARM have no choice but to - unless they wish to concede ground to others). The end result is that we may benefit from faster, better, ARM's in future.
Regards
Annraoi |
guestx
3/8/04 2:19PM |
On the subject of power consumption, doesn't someone need to mention that the screen and hard disk of an ARM-based solution would consume a fair amount of power, too? Nobody is doubting that ARM-based CPUs are very suitable for low power consumption applications (see http://www.gumstix.com for some nice kit), but add a screen and other components and I'd imagine that the benefits start to fade away. |
scl4c0rn
3/8/04 2:49PM |
in reply to nunfetishist:
"What's all this about VRPC funding Windows? I imagine it's such a TINY
number of sales, nobody at Microsoft have noticed"
I don't care whether M$ have noticed or not. It's the principle: why are
we forced to use Windows? Why don't they (STD) offer an alternative?
|
| Use the forum for more comments on this article |
