| |
RON to go ahead |
| |
This is a long thread. Click here to view the threaded list. |
| |
|
Guy Inchbald |
Message #89276, posted at 12:55, 23/10/2001 |
| Unregistered user
|
Hooray!
IOMD? VIDC? How is it done? |
| |
 [ Log in to reply ] |
| |
|
Andrew Weston |
Message #89277, posted at 13:05, 23/10/2001, in reply to message #89276 |
| Unregistered user
|
You just write new sections of code.
This would be the case for any hardware developer that comes along with a new proposal I understand. |
| |
[ Log in to reply ] |
| |
|
Jeffrey Lee |
Message #89278, posted at 13:42, 23/10/2001, in reply to message #89277 |
| Unregistered user
|
Oooh ooh ooh! I want one!
*pulls face at silly keyboardless PDA thingies*
Now all we need is a solar powered version... |
| |
[ Log in to reply ] |
| |
|
Lee Johnston |
Message #89279, posted at 14:38, 23/10/2001, in reply to message #89278 |
| Unregistered user
|
It's not that simple - if it was just a case of writing "new chunks of code" I suspect we'd have seen Evolution by now.
You only have to look at the hoops John Kortnik jumped though to get Viewfinder working (and the admission from Paul Middleton in AU that it's a hack and not supported by the OS) to see it's not simple - and the graphics should be the easy bit as the OS should be using the VDU and Sprite abstractions. I'd expect IOMD to be the really difficult bit.
What concerns me is whether this is a step towards hardware independence or just a case of porting RISC OS to another, minority chipset.
|
| |
[ Log in to reply ] |
| |
|
[mentat] |
Message #89280, posted at 14:45, 23/10/2001, in reply to message #89279 |
| Unregistered user
|
I share your concern completely.
However, either way it's better than not at all, I suppose? |
| |
[ Log in to reply ] |
| |
|
Jason Tribbeck |
Message #89281, posted at 15:03, 23/10/2001, in reply to message #89280 |
| Unregistered user
|
Well, it looks like I'll have to get one then, and port ArcCommand onto it! |
| |
[ Log in to reply ] |
| |
|
Richard Goodwin |
Message #89282, posted at 15:06, 23/10/2001, in reply to message #89281 |
| Unregistered user
|
Better than not at all? Well, that depends if it takes resources to do this for a platform which may not take off, may not be manufactured in a few months, and gives nothing back to the "mainstream" RISC OS development.
On the other hand the RON netBook might take off and introduce new people to RISC OS, and/or the modifications help make RISC OS hardware independant and open the market to a whole range of new machines from PDAs to ninja desktop machines.
I'll reserve judgement on what's more likely, readers will no doubt make up their own minds. Time will tell. I wouldn't mind one though. |
| |
[ Log in to reply ] |
| |
|
Gunnlaugur Jonsson |
Message #89283, posted at 16:45, 23/10/2001, in reply to message #89282 |
| Unregistered user
|
Jason, doesn't the press release say that it "will allow the use of all the current Acorn / RISC OS software". Would you then have to port ArcCommand to it?
If it does allow all current Acorn/RISC OS software to run surely this is excellent news, given that the pricing will be reasonable. Trouble is, this might actually give the RiscStation portable a really hard time.
|
| |
[ Log in to reply ] |
| |
|
Willy Mitchell |
Message #89284, posted at 17:22, 23/10/2001, in reply to message #89283 |
| Unregistered user
|
All the better - the RiskStation laptop is an expensive slow thing that's utterly pointless. |
| |
[ Log in to reply ] |
| |
|
Andrew Weston |
Message #89285, posted at 18:16, 23/10/2001, in reply to message #89284 |
| Unregistered user
|
Well, sections of code have surely been written for this machine because funding has been available to enable the company to pay programmers to do this. At least that's how any alternatives to the Acorn memory and video controllers if present within this machine would have been supported.
I'm sure the Evolution would require a lot more as you say, Lee. |
| |
[ Log in to reply ] |
| |
|
Steve |
Message #89286, posted at 00:43, 24/10/2001, in reply to message #89285 |
| Unregistered user
|
Great, someone has decided to put some money up! I can't see that it can be anything but positive all round.
Hopefully they will be cheap enough for schools to by class sets! |
| |
[ Log in to reply ] |
| |
|
Lee Johnston |
Message #89287, posted at 08:42, 24/10/2001, in reply to message #89286 |
| Unregistered user
|
I can't see how the work on this and Evolution would be any different - you have to remove the VIDC and IOMD dependencies for both. The point I'm tring to make is whether this port will see it being done in controlled fashion (ie proper abstraction) or whether it'll just hack the parts of RISC OS that are problematic, leaving us no better off as far as new, top end, hardware is concerned.
As for porting ArcCommand - I suspect that, at the very least, Jason will have to test and maybe rework the graphics plotting code unless the NetBook just happens to have exactly the same video formats as the VIDC20.
|
| |
[ Log in to reply ] |
| |
|
Andrew Weston |
Message #89288, posted at 09:40, 24/10/2001, in reply to message #89287 |
| Unregistered user
|
What I mean is that the modifications seem to have been made because there has been the backing to make the adjustments. An abstraction has not been made apparently and so this would have to be done for any new hardware which we're told requires money that ROL have not got. |
| |
[ Log in to reply ] |
| |
|
Jason Tribbeck |
Message #89289, posted at 13:34, 24/10/2001, in reply to message #89288 |
| Unregistered user
|
Gunnlaugur - keyboard layout'll be different, and I'll think about reducing the game's size. |
| |
[ Log in to reply ] |
| |
|
Jeffrey Lee |
Message #89290, posted at 17:28, 24/10/2001, in reply to message #89289 |
| Unregistered user
|
Mouse emulation could be a bit of a problem too... |
| |
[ Log in to reply ] |
| |
|
John Hoare |
Message #89291, posted at 17:59, 24/10/2001, in reply to message #89290 |
| Unregistered user
|
Anyone any idea how much it'll cost? I need to start saving :-) |
| |
[ Log in to reply ] |
| |
|
Steve |
Message #89292, posted at 19:12, 24/10/2001, in reply to message #89291 |
| Unregistered user
|
Once RISCOS is modified to work on the SA1100 it opens up the possibility of developing other hardware around that processor. No expensive custom chips required. It may not be as fast as something powered by an XScale, but it could definitely be built for less.
|
| |
[ Log in to reply ] |
| |
|
Richard Goodwin |
Message #89293, posted at 08:32, 25/10/2001, in reply to message #89292 |
| Unregistered user
|
Mouse shouldn't be too much of a problem should it? The touch screen should handle most of it, and a key combo could handle menu. Not much different from this graphics tablet I've got emulating the mouse...? |
| |
[ Log in to reply ] |
| |
|
Lee Johnston |
Message #89294, posted at 09:43, 25/10/2001, in reply to message #89293 |
| Unregistered user
|
Steve - no it won't because RISC OS will still be tied to either the VIDC and IOMD and whatever the netbook uses.
The processor isn't the issue - it's basically a SA core designed to be integrated with other chips so the ARM code in RISC OS will run without a problem. However if RISC OS is just ported to netbook only we still can't use other graphics / IO controller hardware.
|
| |
[ Log in to reply ] |
| |
|
John Duffell |
Message #89295, posted at 11:09, 25/10/2001, in reply to message #89294 |
| Unregistered user
|
Suerly IOMD would be easier, the CPU should be able to access memory the same, it's just altering the mapping and altering the protection levels which would have to be changed? However, I'm not well infomed on this, but I would have thought VIDC would be harder. |
| |
[ Log in to reply ] |
| |
|
Rob Dimond |
Message #89296, posted at 15:47, 25/10/2001, in reply to message #89295 |
| Unregistered user
|
This is the best thing that has happened to RISC OS in the last two years, at least.
IMHO it's better than a long-awaited RISC OS laptop since it takes advantage of the power efficiency of the ARM and the RAM and mass storage efficiency of RISC OS. I want one.
|
| |
[ Log in to reply ] |
| |
|
Lee Johnston |
Message #89297, posted at 15:55, 25/10/2001, in reply to message #89296 |
| Unregistered user
|
AIUI (and correct me if I'm wrong) the problem with IOMD is that, as an IO controller, your OS is inherently tied to how it works. Also I get the impression that the dependencies are not contained in one place in the code.
As far as VIDC goes IF the OS worked entirely through its own graphics abstractions (ie sprite modules and VDU drivers) then all that would be needed would be different implementations that could talk to different hardware. Again the viewfinder shows it's not quite that simple although the fact that viewfinder has been done, whereas MD and Millipede were trying to emulate the IOMD, suggests the removing VIDC dependencies should be easier.
That was also the opinion of Simtec when they first showed the Evolution board.
|
| |
[ Log in to reply ] |
| |
|
Steve Wood |
Message #89298, posted at 11:35, 26/10/2001, in reply to message #89297 |
| Unregistered user
|
Yep, I'm seriously considering one now, but primarily because Java development/ARM are vital criteria for my next purchase(s) - especially now that M$ have abandoned Java on future PC's!
Being able to run my RISC OS stuff on-the-move will be a nice bonus and ppl around me will get to see what RoS can do ;-).
[Pity we don't (won't?) have *real* Java strength on RISC OS itself tho' :-((. IMHO, demand is really hotting up for new 'Java enabled' desktops/laptops. The netbook looks like a serious (business) contender if Psion can market it right! (Sigh!) And Java's a good ploy - its got that mass 'BBC BASIC' curiosity about it, but limited hardware/tools (i.e PC based!) to capitalise on it ATM. A Lack of Java on PC's will soon generate a big hardware/OS opportunity for someone. If not, the outcome will be a lingering death to a .net universe :-(]
I hope RISCOS Ltd do a really great job on this - we need an important breakthrough SOON! |
| |
[ Log in to reply ] |
| |
|
Lee Johnston |
Message #89299, posted at 11:52, 26/10/2001, in reply to message #89298 |
| Unregistered user
|
I don't think it's likely that Java will appear on RISC OS and I certainly can't see how anyone could justify the effort involved. It gets harder with every release and the fact that RISC OS is seriously lacking in certain areas doesn't help.
As for MS not supporting Java on Windows...so what? Sun are. I've even heard rumours of Java being ported to the .net framework. MS aren't the be all and end all of development tools on Windows.
|
| |
[ Log in to reply ] |
| |
|
John Duffell |
Message #89300, posted at 11:47, 28/10/2001, in reply to message #89299 |
| Unregistered user
|
Lee: What do you mean "inherently"? Surely the VIDC is just as inherent? Things like the PS2 driver are one module, then you have a serial block driver, the parallel module, then stuff in the kernel for mapping pages and task swapping and things. You also have floppy drives and things. VIDC however, is in things like the font manager, spriteextend, the kernel (for setting stuff and plotting sprites) seems equally spread out to me. Then of course youhave things writing directly to the screen.
BFN |
| |
[ Log in to reply ] |
| |
|
Lee Johnston |
Message #89301, posted at 13:39, 28/10/2001, in reply to message #89300 |
| Unregistered user
|
I think you've made my point for me there - if you're right then VIDC specific functionality is spread across RISC OS and therefore RISC OS is inherently tied to the VIDC.
If there is sprite plotting code in the kernel then the people who put it there should be shot...
As for applications writing directly to the screen I don't see the issue with this. The developers chose to bypass the RISC OS APIs, mainly for performance reasons, so if moving to different hardware breaks them then tough. Certainly I don't think it's RISC OS Ltds job to try and ensure such applications run on new hardware and I don't think the development of new hardware should be stifled by such considerations. I should point out that I have applications that this would break (including games I've written myself).
OS' like Windows and Linux specify an abstraction of the video processor (the HAL if you like) and then leave it to drivers to implement this abstraction. The parts of the OS which require access to the screen talk to it through this abstraction.
As for games with the advent of DirectX on Windows there has been less and less need to talk directly to the hardware. I'd be surprised if any games still did. You can do it if you want but it's a lot of work.
|
| |
[ Log in to reply ] |
| |
|
Steve Wood |
Message #89302, posted at 00:58, 2/11/2001, in reply to message #89301 |
| Unregistered user
|
Lee, re: Java stuff above . . .
I'm not sure I made my point clearly - I was trying to say that alot of ppl working with Java applications (particularly database developers like me!) just want some OS that works. EPOC already has several Java alternatives and SOMEONE is going to make a killing on the OS/Hardware side when it comes to both the development and running of Java apps. If RISC OS can't do it, then IMHO this will be a serious restriction - helping to prevent the growth of the system that we all need and want [a real missed opportunity to attract and hold onto new users :-( ] Java is obviously important to ARM judging by their licensing and other efforts with Jazelle. Anyone got any ideas how/whether Jazelle could be used alongside RISC OS to provide Java functionality (or not) ?? |
| |
[ Log in to reply ] |
| |
|
Lee Johnston |
Message #89303, posted at 10:05, 2/11/2001, in reply to message #89302 |
| Unregistered user
|
Steve - I understand exactly where you're coming from. I was just making the point that porting Java is a huge effort, and it gets harder all the time as newer versions of Java begin to move away from the "lowest common denominator" ethic of the early versions to "utilise the features of mainstream OS'" that we're starting to see in version 1.4.
Besides the complexity of porting all the AWT and Swing libraries RISC OS has other failings - no multithreading as standard means that it would have to be created. The lack of pre-emptive multitasking would make implementing the new I/O libraries difficult - developers on other platforms would (quite rightly IMHO) expect to be able to block an application on the new channel selectors without blocking the whole OS. I imagine that under RISC OS this would require some trickery whereby Java apps are WIMP tasks and channel events are passed through the WIMP events and then dispatched to where the Java app expects them. Alternatively if the threading scheme implemented is pre-emptive then that could be exploited. Either way it's not exactly trivial.
As for Jazelle - that would help with the implementation of the virtual machine, not the libraries. Again as people like Peter Naulls have pointed out, getting the VM running is probably the easy bit. Mind you this is a bit academic because RISC OS can't run on the ARM processors which will feature Jazelle (ARM9 upwards).
|
| |
[ Log in to reply ] |
| |
|
Jeffrey Lee |
Message #89304, posted at 15:51, 2/11/2001, in reply to message #89303 |
| Unregistered user
|
.... Until Omega ....
*Sits and waits*
|
| |
[ Log in to reply ] |
| |
|
Lee Johnston |
Message #89305, posted at 16:49, 2/11/2001, in reply to message #89304 |
| Unregistered user
|
Omega won't run RISC OS on an ARM9+ processor either. AIUI it'll run applications compiled to run on those processors by some switching.
In this case I guess a JVM could be written and compiled to 32bit address mode and then the Omega could switch it onto the Jazelle enabled processor which could execute the Java bytecode.
Two problems
1) MD are aiming at XScale (no Jazelle)
2) You still don't have any of the Java libraries and a port still has to overcome the "deficencies" in RISC OS.
|
| |
[ Log in to reply ] |
| |
| Pages (2): 1
> >|
|
RON - the RISC OS on netBook project - has secured funding, on the back of news that the netBook will continue to be produced despite Psion's recent jitters over consumer hardware.