I don't think it's supposed to be competing with the PSP or DS, to be honest. I'm a big fan of the GP32 (what came before), and I'm further an owner of one of these (and it is a great deal cooler as devices go, despite the loopholes), but there's no way I'd claim this thing can compete with either. The closest it'll come graphically is a decent PSX emulator (which is already chugging along somewhat, although needs some more fiddling, if the last vid I saw of it was anything to go by), or, conceivably, a port of the wonderful Yeti3D engine - if and only if people actually start to use it!
What it *will* do is embrace homebrew in a way that the PSP or DS are unlikely ever to do, outside of an incredibly restricted box. For those that want graphical splendour, therefore, it's of little importance - buy a PSP and be wowed by the graphics... that's the main thing about the thing. On the other hand, for those that want the buzz of a truly thriving community that's releasing new amazing developments every few days, both in emulators, but also in some very addictive homebrew games (Tilematch springs to mind), then this is pretty much the only thing to go for.
Unless, of course, you want to go for a Zaurus or what have you instead, which, if we're honest, has somewhat more punch (albeit a larger price tag too), but at the expense of... well, having buttons in the right places to play stuff, if all the reports I've heard are in fact true!
To be honest, if Lindows had trouble with MS over a single letter shift, I'd be incredibly surprised if you could get away with a similar shift for an even more similar product, albeit with a company a LOT smaller than MS being the pissed off people.
See, I am quite certain that, were my mobile to offer a screen of at least the same size as a GBA and free TV capability, then I would watch it. I've certainly used my GP32 to watch movies on for long journeys these last couple of years, despite (or to some extent, because of - it IS usable on a bus as a TV isn't) the 3 1/2 inch screen.
However, its worth remembering that if money can be milked from something (especially by phone companies) it will be... SMSes are still, iirc, ridiculously overpriced, just as are picture messages (can't remember the figures, but a few years ago it was something like 0.01p cost and 10p being charged by the phone companies), and were a mobile phone to tout its video capabilities, I'd automatically assumes I'd be paying through the nose to get at them.
Which I imagine most are unwilling to do. The question should be wether eople would be interested in *free* TV on their mobile handsets, or whether, say, you would choose a handset with 3 days' battery life with TV against one with 20 days' battery life without. Or whatever other features you want to compare with TV's desirability.
Well, you'll find the stats to some extent at www.gp2x.com. At least if its working. Certainly they were there for a long while.
Roughly speaking, the chip is the MMSP2 one from MagicEyes. That means it has an Arm920t core and an Arm940t one both at 200MHz. Also has a few other controllers for various crizzle but I can't remember them all.
Uses SD cards, has 64Mb RAM proper, and 64Mb NAND flash memory (which is still open debate over whether its either possible or sensible to store stuff in on account of likely limited writes) despite being perfectly fast enough to use as ROM storage.
USB2 is, I think, Unpowered Host. If it does host, of course, which I'm fairly sure it does; just not with any power being supplied. If your devices are externally powered, this should still be OK, I imagine.
No idea about ebooks, sorry.
The SDK on release is said to be SDL. I assume there will be some extensions to allow better use of the second core, but I honestly don't know - but SDL is the primary thing which people'll be starting with. IIRC a wonderful guy called Mr_Mirko is also going to be making his own SDK (can't be open because MagicEyes insist on coders signing an NDA before giving them access to the lower level stuff), which I would hope will definitely be designed with 2 cores in mind. He's made a wonderful SDK for the older GP32 (I was actually able to do more then "Hello World" with it - I'm not a good coder lol). Of course, it isn't made yet, so kinda hard to give it to you:P.
Hope that's helped a little
-- Toby
Aside from Linux gaming (which I suspect comprises quite a whack of the conversation - I'm afraid I haven't bothered to read it, it's just too late:P ), and after we got over the whole psychological need for emulators and ports, it seems to have worked on almost precisely that suggested model.
There is very little very glitzy on the machine, but there's some stunning native and original puzzle games that haven't been done before, as well as not-quite-commercial-quality games like Nazcadreams, Nazcarunners (both using Fenix, so you'll be able to play them on a PC unless I'm much mistaken with no trouble) or even GPSpout. Once upon a time the games for it were limited seriously to what you could emulate, but these days, there really is quite a lot. The guys who make it generally get a decent ammount of respect for their work and (albeit not so often) from time to time donations from people in the community. And of course, they absolutely control their own code; they wrote it in all likelyhood from a bedroom for fun, notably without needing a license from somewhere to do so, so there's absolutely no reason why their IP ends up in the hands of a corporate giant, although conceivably it might at some stage end up in the public domain. If they say so.
True, such a machine doesn't leave a great deal of space left for commercial development simply because there's so much available for free (whether legally or less so in the form of emulable roms) that is of commercial quality (in the case of emulated stuff because it once was, of course). But considering the hardware itself was profitable it didn't have to be.
Ok this is a handheld, not a PC, and the situation is entirely different on the two, principally because of the lack of the culture of hardware being subsidized and homebrew being made as hard as possible in desktop computing which exists in the gaming handheld industry, alongside a great deal of other factors that for some reason aren't springing to mind. But that said, its still worth pointing out as an instance where the idea was at least partially successful.
True. On the other hand, I don't quite see why it wouldn't be possible (if extremely long) to tell a program what it was looking for ("Hi there 123") and then tell it to perform every conceivable operation on the jumbled mess ("Jk vjgtg 345") filtering off those operations that resulted in the original message. Clearly, the shorter the message, the more operations this would be (proportionally you know what less of the data is), but among them, -2 would exist; likewise -3 for your chip.
Once you've found a set for one email, feed it another, and another, narrowing down the search list each time until you've got one result - your own master key, as it were.
Of course, this would take a long time - especially since the larger the key, the more complext the solution can be. However, I doubt trusted chips *could* have keys greater than a certain length if only because the larger the key, the larger the encrypted file - clearly, a key of 1gb length is going to result in some ridiculously large file for even the simplest of messages. However, they can, as such, not be used because no consumer would touch a 40meg email with a bargepole - especially if it had to be decrypted to see if it was simply an ad for viagra from someone they'd never met. Let alone what ISPs would think. It'd be self-defeating as a system.
Essentially, I don't believe it would ever be unbreakable - just take a very long time to break.
Assuming of course that it would not be possible to use the method described briefly at the Register a few months back - pre-calc tables, I think, essentially - to do it faster. Which might well be necessary of course - after all, there would be little point, I suppose, in cracking a trust chip using your entire CPU power over a period like 20 years. It'd be obsolete by the time you finished.
Out of interest, would it be remotely possible to gain your master key through the use of trusted emails, perchance? Or something similar at least.
I mean, if I'm getting it right, when a piece of trusted software comes through, you have an excrypted lump of goo sitting on your HDD. Then you run it through the trust chip and it essentially cleans it up until its usable (perhaps a bad description considering what's really going on).
However.
The enigma machine code was broken by knowing what was supposed to be being said. The weather was obvious. It was encoded using the Enigma. Therefore if you have the weather code book and the encoded message you can decode all messages.
So. Think about it. Send yourself an email - preferably a long one. Make sure it is trusted. You, as the author of that email, know precisely what's in it. If you send it via a local private SMTP server directly to an account on the internal network, then you probably know all the header information too. And you've also got the "trusted" email, all nicely encrypted up. Do that a few times, and surely it would be possible (if not easy) to create a program to decode the trusted email into the original one. Which is, presumably, your master key.
Voila. Broken trusted computing.
Any chance anyone could tell me why this wouldn't work?
Surely this would not be intensely hard to get round. It needn't, after all, be one certification per distro - one cert for a single bootloader app ought to be enough surely? Once you've got something other than the default Windows loader allowing you to choose where to go next, its just a matter of tweaking - e.g. including an option to emulate a pre-trust bios in that bootloader (which is certified as being trusted) and boot from a CD, ISO, or wheresoever you want.
Ok, the first certification process would be expensive. Bloody expensive. But lets face it, we're a community of 20000 plus here (hence the term Slashdotting). If every one of us donated £5, that's £100000. If 20% of us did, that's still £20000 - and I cannot believe the fee would be much more than that; after a while, it starts to bite into even the big companies. Once there's one properly designed free program that the bios knows it can trust, then free software has a way in, surely. And once there's a way in, the plan has, essentially, failed.
At least for M$.
£5 for free computing for the rest of this horrible era we're moving into. For me that's worth it.
Granted, this way, the entertainment industry still has a lot of control. At least, until someone developes a trust chip emulator (by trial and error most likely, but I'd be surprised if it didn't happen). Then, of course, we suddenly find ourselves in a somewhat more favourable position than we do at the moment. Sacrifice some CPU time to emulate it, create a virtual virtual sandbox that gets there first, and all of a sudden the industry thinks it can trust you but no longer can.
We, gentlemen, are on the brink of a long and bloody war. But its by no means one - even should trusted computing become the standard that everyone expects - that we cannot win.
Re:Phantom haptic devices
on
3D Mouse
·
· Score: 1
At a very brief glance that's pretty close to what I was on about when I was talking about the arm made out of lego mixed with a ripped apart mouse and a digicam. Doesn't surprise me someone's already made a nice-looking non-hack version though;)
I must say, I don't quite see why you couldn't adapt a standard 2D optical mouse for 3D uses simply by making the sensor more sensitive and analogue (i.e. not just returning 1 or 0 for the strength of the light, but also how strong the light is). Then you just need a plain mousemat, and some drivers to read the newly put in sensor, and you should be able to track the mouse's Z co-ordinate. Maybe.
Or perhaps using a webcam would work a little better. Constant pattern and the tracing of the images of that pattern would give direction in X-Y, and the size of the pattern would give the Z-co-ord. Problem there would be the lack of buttons, of course, but I don't imagine that'd be one *too* hard to deal with... you could always stick a digicam to the bottom of a mouse and use that for motion detecting, and the mouse-buttons for actual clicks.
There's probably quite a few other ways to do a similar trick as well...
Perhaps the best (though certainly the most complex) would be a 3D arm with a cam on the end and a couple of buttons. And probably a release button to be held to move in Z so you could keep your hand rested on it. I imagine you could make one without too much trouble, though, out of Lego Technic, a few bits of mouse, and a digicam. And possibly a small light to make sure the mat was always illuminated enough.
Anyways.
Slashdot Mirror
I don't think it's supposed to be competing with the PSP or DS, to be honest. I'm a big fan of the GP32 (what came before), and I'm further an owner of one of these (and it is a great deal cooler as devices go, despite the loopholes), but there's no way I'd claim this thing can compete with either. The closest it'll come graphically is a decent PSX emulator (which is already chugging along somewhat, although needs some more fiddling, if the last vid I saw of it was anything to go by), or, conceivably, a port of the wonderful Yeti3D engine - if and only if people actually start to use it!
What it *will* do is embrace homebrew in a way that the PSP or DS are unlikely ever to do, outside of an incredibly restricted box. For those that want graphical splendour, therefore, it's of little importance - buy a PSP and be wowed by the graphics... that's the main thing about the thing. On the other hand, for those that want the buzz of a truly thriving community that's releasing new amazing developments every few days, both in emulators, but also in some very addictive homebrew games (Tilematch springs to mind), then this is pretty much the only thing to go for.
Unless, of course, you want to go for a Zaurus or what have you instead, which, if we're honest, has somewhat more punch (albeit a larger price tag too), but at the expense of... well, having buttons in the right places to play stuff, if all the reports I've heard are in fact true!
To be honest, if Lindows had trouble with MS over a single letter shift, I'd be incredibly surprised if you could get away with a similar shift for an even more similar product, albeit with a company a LOT smaller than MS being the pissed off people.
See, I am quite certain that, were my mobile to offer a screen of at least the same size as a GBA and free TV capability, then I would watch it. I've certainly used my GP32 to watch movies on for long journeys these last couple of years, despite (or to some extent, because of - it IS usable on a bus as a TV isn't) the 3 1/2 inch screen. However, its worth remembering that if money can be milked from something (especially by phone companies) it will be... SMSes are still, iirc, ridiculously overpriced, just as are picture messages (can't remember the figures, but a few years ago it was something like 0.01p cost and 10p being charged by the phone companies), and were a mobile phone to tout its video capabilities, I'd automatically assumes I'd be paying through the nose to get at them. Which I imagine most are unwilling to do. The question should be wether eople would be interested in *free* TV on their mobile handsets, or whether, say, you would choose a handset with 3 days' battery life with TV against one with 20 days' battery life without. Or whatever other features you want to compare with TV's desirability.
Well, you'll find the stats to some extent at www.gp2x.com. At least if its working. Certainly they were there for a long while. Roughly speaking, the chip is the MMSP2 one from MagicEyes. That means it has an Arm920t core and an Arm940t one both at 200MHz. Also has a few other controllers for various crizzle but I can't remember them all. Uses SD cards, has 64Mb RAM proper, and 64Mb NAND flash memory (which is still open debate over whether its either possible or sensible to store stuff in on account of likely limited writes) despite being perfectly fast enough to use as ROM storage. USB2 is, I think, Unpowered Host. If it does host, of course, which I'm fairly sure it does; just not with any power being supplied. If your devices are externally powered, this should still be OK, I imagine. No idea about ebooks, sorry. The SDK on release is said to be SDL. I assume there will be some extensions to allow better use of the second core, but I honestly don't know - but SDL is the primary thing which people'll be starting with. IIRC a wonderful guy called Mr_Mirko is also going to be making his own SDK (can't be open because MagicEyes insist on coders signing an NDA before giving them access to the lower level stuff), which I would hope will definitely be designed with 2 cores in mind. He's made a wonderful SDK for the older GP32 (I was actually able to do more then "Hello World" with it - I'm not a good coder lol). Of course, it isn't made yet, so kinda hard to give it to you :P.
Hope that's helped a little
-- Toby
Aside from Linux gaming (which I suspect comprises quite a whack of the conversation - I'm afraid I haven't bothered to read it, it's just too late :P ), and after we got over the whole psychological need for emulators and ports, it seems to have worked on almost precisely that suggested model.
There is very little very glitzy on the machine, but there's some stunning native and original puzzle games that haven't been done before, as well as not-quite-commercial-quality games like Nazcadreams, Nazcarunners (both using Fenix, so you'll be able to play them on a PC unless I'm much mistaken with no trouble) or even GPSpout. Once upon a time the games for it were limited seriously to what you could emulate, but these days, there really is quite a lot. The guys who make it generally get a decent ammount of respect for their work and (albeit not so often) from time to time donations from people in the community. And of course, they absolutely control their own code; they wrote it in all likelyhood from a bedroom for fun, notably without needing a license from somewhere to do so, so there's absolutely no reason why their IP ends up in the hands of a corporate giant, although conceivably it might at some stage end up in the public domain. If they say so.
True, such a machine doesn't leave a great deal of space left for commercial development simply because there's so much available for free (whether legally or less so in the form of emulable roms) that is of commercial quality (in the case of emulated stuff because it once was, of course). But considering the hardware itself was profitable it didn't have to be.
Ok this is a handheld, not a PC, and the situation is entirely different on the two, principally because of the lack of the culture of hardware being subsidized and homebrew being made as hard as possible in desktop computing which exists in the gaming handheld industry, alongside a great deal of other factors that for some reason aren't springing to mind. But that said, its still worth pointing out as an instance where the idea was at least partially successful.
True. On the other hand, I don't quite see why it wouldn't be possible (if extremely long) to tell a program what it was looking for ("Hi there 123") and then tell it to perform every conceivable operation on the jumbled mess ("Jk vjgtg 345") filtering off those operations that resulted in the original message. Clearly, the shorter the message, the more operations this would be (proportionally you know what less of the data is), but among them, -2 would exist; likewise -3 for your chip.
Once you've found a set for one email, feed it another, and another, narrowing down the search list each time until you've got one result - your own master key, as it were.
Of course, this would take a long time - especially since the larger the key, the more complext the solution can be. However, I doubt trusted chips *could* have keys greater than a certain length if only because the larger the key, the larger the encrypted file - clearly, a key of 1gb length is going to result in some ridiculously large file for even the simplest of messages. However, they can, as such, not be used because no consumer would touch a 40meg email with a bargepole - especially if it had to be decrypted to see if it was simply an ad for viagra from someone they'd never met. Let alone what ISPs would think. It'd be self-defeating as a system.
Essentially, I don't believe it would ever be unbreakable - just take a very long time to break.
Assuming of course that it would not be possible to use the method described briefly at the Register a few months back - pre-calc tables, I think, essentially - to do it faster. Which might well be necessary of course - after all, there would be little point, I suppose, in cracking a trust chip using your entire CPU power over a period like 20 years. It'd be obsolete by the time you finished.
Out of interest, would it be remotely possible to gain your master key through the use of trusted emails, perchance? Or something similar at least.
I mean, if I'm getting it right, when a piece of trusted software comes through, you have an excrypted lump of goo sitting on your HDD. Then you run it through the trust chip and it essentially cleans it up until its usable (perhaps a bad description considering what's really going on).
However.
The enigma machine code was broken by knowing what was supposed to be being said. The weather was obvious. It was encoded using the Enigma. Therefore if you have the weather code book and the encoded message you can decode all messages.
So. Think about it. Send yourself an email - preferably a long one. Make sure it is trusted. You, as the author of that email, know precisely what's in it. If you send it via a local private SMTP server directly to an account on the internal network, then you probably know all the header information too. And you've also got the "trusted" email, all nicely encrypted up. Do that a few times, and surely it would be possible (if not easy) to create a program to decode the trusted email into the original one. Which is, presumably, your master key.
Voila. Broken trusted computing.
Any chance anyone could tell me why this wouldn't work?
Surely this would not be intensely hard to get round. It needn't, after all, be one certification per distro - one cert for a single bootloader app ought to be enough surely? Once you've got something other than the default Windows loader allowing you to choose where to go next, its just a matter of tweaking - e.g. including an option to emulate a pre-trust bios in that bootloader (which is certified as being trusted) and boot from a CD, ISO, or wheresoever you want. Ok, the first certification process would be expensive. Bloody expensive. But lets face it, we're a community of 20000 plus here (hence the term Slashdotting). If every one of us donated £5, that's £100000. If 20% of us did, that's still £20000 - and I cannot believe the fee would be much more than that; after a while, it starts to bite into even the big companies. Once there's one properly designed free program that the bios knows it can trust, then free software has a way in, surely. And once there's a way in, the plan has, essentially, failed. At least for M$. £5 for free computing for the rest of this horrible era we're moving into. For me that's worth it. Granted, this way, the entertainment industry still has a lot of control. At least, until someone developes a trust chip emulator (by trial and error most likely, but I'd be surprised if it didn't happen). Then, of course, we suddenly find ourselves in a somewhat more favourable position than we do at the moment. Sacrifice some CPU time to emulate it, create a virtual virtual sandbox that gets there first, and all of a sudden the industry thinks it can trust you but no longer can. We, gentlemen, are on the brink of a long and bloody war. But its by no means one - even should trusted computing become the standard that everyone expects - that we cannot win.
At a very brief glance that's pretty close to what I was on about when I was talking about the arm made out of lego mixed with a ripped apart mouse and a digicam. Doesn't surprise me someone's already made a nice-looking non-hack version though ;)
I must say, I don't quite see why you couldn't adapt a standard 2D optical mouse for 3D uses simply by making the sensor more sensitive and analogue (i.e. not just returning 1 or 0 for the strength of the light, but also how strong the light is). Then you just need a plain mousemat, and some drivers to read the newly put in sensor, and you should be able to track the mouse's Z co-ordinate. Maybe. Or perhaps using a webcam would work a little better. Constant pattern and the tracing of the images of that pattern would give direction in X-Y, and the size of the pattern would give the Z-co-ord. Problem there would be the lack of buttons, of course, but I don't imagine that'd be one *too* hard to deal with... you could always stick a digicam to the bottom of a mouse and use that for motion detecting, and the mouse-buttons for actual clicks. There's probably quite a few other ways to do a similar trick as well... Perhaps the best (though certainly the most complex) would be a 3D arm with a cam on the end and a couple of buttons. And probably a release button to be held to move in Z so you could keep your hand rested on it. I imagine you could make one without too much trouble, though, out of Lego Technic, a few bits of mouse, and a digicam. And possibly a small light to make sure the mat was always illuminated enough. Anyways.