Well Acorn invented the BBC micro but the BBC's efforts to educate the populous about the 'micro revolution' in the early eighties are a fine example of how the BBC can actually help the country.
On the contrary. I live in the UK and fully support the license fee. The BBC is one of the best things about this country, even if the quality of their news output has dropped of late.
In any case, actual trivial embedding code is the best documentation IMHO. I detect an attempt at personal abuse suggesting that I'm not a professional developer and would know nothing in this regard. It isn't worth getting into an e-penis size discussion in this regard but if you don't view the wealth of Gecko stuff available as enough then you've no business calling yourself a programmer.
Thats all very well, but the article isn't about Wikipedia so much as the BBC. It was the Beeb that put up the fake article about a fake dead pop star.
No it wasn't.
It was also a BBC man (from their own network IP range) that put up the fake Boy*Up (?) article too. Although he says he acted alone and not on behalf of the BBC, what are the chances of a BBC man putting up an article connected to a fake BBC website coincidentally?
Given the number of things I come across form the BBC on a daily basis I'd say quite high. Besides he gave a plausable method of hitting it. If you really don't believe him why not ask him about it on his talk page?
Firstly because they're not a company? Secondly because they are not a news corporation - they are a broadcasting corporation (the name kinda gives it away). Thirdly because the original author of the article was just some student somewhere with nothing to do with the BBC at all. Fourthly because the damn game has been running for ages so this is a bit of a piss poor time to market it. Fifthly, the BBC would get far more milage out of advertising it on its won damn site than on a random Wikipedia page.
That's fine, but then this isn't a kernel component any longer.
Correct. In an exactly analagous manner the DRM isn't a kernel component but the hardware tasks 'required' are. The point of a stack is that the APIs between layers are well defined and that propriety components can be layered in. Your entire computer is a stack with the GPL-d kernel sitting between the BIOS (most likely propriety) and, e.g., Acrobat Reader (propriety).
The thing to come away with here is that just because the hardware driver for TPM is in the kernel it doesn't mean that a userspace DRM implementation which uses the driver needs to be GPL.
But we're not talking about crypto signing here, we're talking about DRM.
You are spectacularly uninformed about the bread and butter of DRM:).
Your point appears to be that any hardware device can always be emulated and such an emulated device can be dropped into a GPL kernel to pass whatever important bits of info might be required back to userspace. This is of course true to first order. The problem with the trusted computing model is that the whole stack is involved.
In order to emulate the entirety of a TPM one has to be able to emulate the BIOS's interaction with the chip (roughly equivalent to extracting its 'private key') as well as, possibly, the kernel or perhaps even the DRM app itself. The Trusted Computing platform was designed explicitly to involve all levels in the modern machine stack even below the kernel so that the entire stack has to be replaced before true emulation is possible.
The problem with TCPA is not that it is particularly hard to emulate 'enough' but that your own hardware can betray you. Suppose you emulate a TPM so that all your userspace applications see your software-chip. All fine and dandy but what about your latest new TPM aware graphics card? It requires that all video above 320x240 requires a TPM signature to display. Unfortunately your graphics card talks directly to the hardware TPM chip, not the kernel's. Your emulated TPM now becomes worthless.
Well... that depends on how Linus is feeling about the GPL that week. Linus has said that certain kinds of non-GPL kernel components are allowed, but it's not always clear what those are
But we're talking about a stack here. Extending your point it would appear that Linux can decide what packets you can send over your network - i.e. you cannot download any non GPL applications! The point here is that a driver can exist irrespective of what use is made of it. You can use all of Linux's drivers to make and distribute non-free content.
In the case of networking, Linux gives you a userspace API for directly modifying the packets sent over the network. You can use that API for whatever non-free things you want (Linux explicitly excludes userspace from the GPL in its license).
My point is that if you have the source to the TPM driver...the DRM would not be any stronger, when you come down to it, than the same code running on a system where the DRM was entirely in user-space.
Except that, to take the case of crypto signing, the private key never leaves the chip so one cannot, by any software means, get access to it. A pure-software implementation would have to contain the private key somewhere.
Think of it like the difference between passive and active smart cards. A passive smart card only contains data and so can be copied by simply reading all the data and burning it into a new card. An active card, OTOH, has the ability to do crypto with some private key locked into the card and which would require physical disassembly and examination under a microscope to extract and copy. This is stronger since an attacker must physically obtain the card and break it in order to copy it.
No system will ever be perfect. It is just a case of creating enough layers to deter enough people.
Now... since this thing operates on kernel level, and if Darwin (the OSX kernel) is OpenSource, is there any chance that the source for DRM stuff will be released as OpenSource too?
What does it mean to have TPM support in a GPL-ed kernel?
Pretty much what it means in a non-GPL kernel. TPM != DRM in much the same way as Ethernet != TCP/IP. One could have a propriety networking stack working over Ethernet just as easily on Linux as on Windows.
What TPM can do is something like 'am I running on the same kernel as I was compiled on' or 'is this version of Linux using this kernel' or 'sign this bit of data and squirrel it away so only I can get to it'.
All of the DRM stuff would be built on top of the TPM driver, not part of it. What would be interesting is a 'dummy' TPM driver which can be controlled from user-space. Then one could see what is going on:).
You sir are a troll who either is incapable of recognising sarcasm or incapable of noting that supporting a piece of hardware doesn't mandate one use it:).
It is the leaders here in England that are justifying. The anti-war sentiment is pretty strong. Unfortunately the propaganda service which pretends to be the US media probably doesn't report this:).
And here, I'll be popping down later with my screwdriver set, breakout box and laptop to see if I can give them a reprogramming they'll never forget...
Since the two postscripts are different, changing the same letter in both will NOT result in the same hash.
Actually the textual part of both is the same, they just found some PS magic to choose between two 'sub-documents' in the same PS. Indeed one can trivially show that modifying both files will result in a collision since both files contain the text of both messages.
OTOH, in the real world, one would be hard pressed to change the contents of an existing document.
Slashdot Mirror
Well Acorn invented the BBC micro but the BBC's efforts to educate the populous about the 'micro revolution' in the early eighties are a fine example of how the BBC can actually help the country.
On the contrary. I live in the UK and fully support the license fee. The BBC is one of the best things about this country, even if the quality of their news output has dropped of late.
This being a UK story, would they use the ATM Machine at the Trustee Savings Bank Bank?
I doubt it. They'd probably just use the cashpoint instead.
I love the way the SSL cert for that site isn't trusted by Firefox.
What about things like icons in a GUI? If a logo is distributed with software as an icon in a GUI, wouldn't that be considered a trademark
It depends if it is a trademark. Unlike copyright trademarks don't happen 'by magic'. Just creating a logo wont make it trademarked.
In no way as far as I can see. How would it?
It's the first guess for finding an inverse sqare root
:= x * x
/-bots, yes I knew what the original poster meant...].
I've got an even faster one: x
[Note to
Maybe an unemployed university student has time to play with such examples that lack documentation. Professional developers do not.
P Is.html
e rview/EmbeddingBasicsTOC.html
t erfaceFreeze.html
That is a bit of a contradiction in terms. One can be unemployed or be employed by a University to carry out research towards a PhD but not both.
In any case here are some design docs.
http://www.mozilla.org/projects/embedding/PublicA
A simple introduction
http://www.mozilla.org/projects/embedding/embedov
The status of the API freeze (i.e. what APIs you can rely on not to change).
http://www.mozilla.org/projects/embedding/EmbedIn
In any case, actual trivial embedding code is the best documentation IMHO. I detect an attempt at personal abuse suggesting that I'm not a professional developer and would know nothing in this regard. It isn't worth getting into an e-penis size discussion in this regard but if you don't view the wealth of Gecko stuff available as enough then you've no business calling yourself a programmer.
http://lxr.mozilla.org/seamonkey/source/embedding/
Use the source luke. Examples of Gecko embedded in Gtk, Win32, Cocoa on OS X and lots more. Even OS/2.
Thats all very well, but the article isn't about Wikipedia so much as the BBC. It was the Beeb that put up the fake article about a fake dead pop star.
No it wasn't.
It was also a BBC man (from their own network IP range) that put up the fake Boy*Up (?) article too. Although he says he acted alone and not on behalf of the BBC, what are the chances of a BBC man putting up an article connected to a fake BBC website coincidentally?
Given the number of things I come across form the BBC on a daily basis I'd say quite high. Besides he gave a plausable method of hitting it. If you really don't believe him why not ask him about it on his talk page?
Firstly because they're not a company? Secondly because they are not a news corporation - they are a broadcasting corporation (the name kinda gives it away). Thirdly because the original author of the article was just some student somewhere with nothing to do with the BBC at all. Fourthly because the damn game has been running for ages so this is a bit of a piss poor time to market it. Fifthly, the BBC would get far more milage out of advertising it on its won damn site than on a random Wikipedia page.
'Asshole' (as I believe you 'merkins say).
That's fine, but then this isn't a kernel component any longer.
:).
Correct. In an exactly analagous manner the DRM isn't a kernel component but the hardware tasks 'required' are. The point of a stack is that the APIs between layers are well defined and that propriety components can be layered in. Your entire computer is a stack with the GPL-d kernel sitting between the BIOS (most likely propriety) and, e.g., Acrobat Reader (propriety).
The thing to come away with here is that just because the hardware driver for TPM is in the kernel it doesn't mean that a userspace DRM implementation which uses the driver needs to be GPL.
But we're not talking about crypto signing here, we're talking about DRM.
You are spectacularly uninformed about the bread and butter of DRM
Your point appears to be that any hardware device can always be emulated and such an emulated device can be dropped into a GPL kernel to pass whatever important bits of info might be required back to userspace. This is of course true to first order. The problem with the trusted computing model is that the whole stack is involved.
In order to emulate the entirety of a TPM one has to be able to emulate the BIOS's interaction with the chip (roughly equivalent to extracting its 'private key') as well as, possibly, the kernel or perhaps even the DRM app itself. The Trusted Computing platform was designed explicitly to involve all levels in the modern machine stack even below the kernel so that the entire stack has to be replaced before true emulation is possible.
The problem with TCPA is not that it is particularly hard to emulate 'enough' but that your own hardware can betray you. Suppose you emulate a TPM so that all your userspace applications see your software-chip. All fine and dandy but what about your latest new TPM aware graphics card? It requires that all video above 320x240 requires a TPM signature to display. Unfortunately your graphics card talks directly to the hardware TPM chip, not the kernel's. Your emulated TPM now becomes worthless.
You're assuming the cylinder was solid. TFA states that it was a cylindrical shell. You need to factor in the thickness of the shell.
Well... that depends on how Linus is feeling about the GPL that week. Linus has said that certain kinds of non-GPL kernel components are allowed, but it's not always clear what those are
But we're talking about a stack here. Extending your point it would appear that Linux can decide what packets you can send over your network - i.e. you cannot download any non GPL applications! The point here is that a driver can exist irrespective of what use is made of it. You can use all of Linux's drivers to make and distribute non-free content.
In the case of networking, Linux gives you a userspace API for directly modifying the packets sent over the network. You can use that API for whatever non-free things you want (Linux explicitly excludes userspace from the GPL in its license).
My point is that if you have the source to the TPM driver...the DRM would not be any stronger, when you come down to it, than the same code running on a system where the DRM was entirely in user-space.
Except that, to take the case of crypto signing, the private key never leaves the chip so one cannot, by any software means, get access to it. A pure-software implementation would have to contain the private key somewhere.
Think of it like the difference between passive and active smart cards. A passive smart card only contains data and so can be copied by simply reading all the data and burning it into a new card. An active card, OTOH, has the ability to do crypto with some private key locked into the card and which would require physical disassembly and examination under a microscope to extract and copy. This is stronger since an attacker must physically obtain the card and break it in order to copy it.
No system will ever be perfect. It is just a case of creating enough layers to deter enough people.
Now... since this thing operates on kernel level, and if Darwin (the OSX kernel) is OpenSource, is there any chance that the source for DRM stuff will be released as OpenSource too?
Probably, Linux's Infineon driver is.
What does it mean to have TPM support in a GPL-ed kernel?
:).
Pretty much what it means in a non-GPL kernel. TPM != DRM in much the same way as Ethernet != TCP/IP. One could have a propriety networking stack working over Ethernet just as easily on Linux as on Windows.
What TPM can do is something like 'am I running on the same kernel as I was compiled on' or 'is this version of Linux using this kernel' or 'sign this bit of data and squirrel it away so only I can get to it'.
All of the DRM stuff would be built on top of the TPM driver, not part of it. What would be interesting is a 'dummy' TPM driver which can be controlled from user-space. Then one could see what is going on
You sir are a troll who either is incapable of recognising sarcasm or incapable of noting that supporting a piece of hardware doesn't mandate one use it :).
OS X has an Infinium driver == evil.
Linus has an Infinum driver == good.
I see. I must be new here.
This is terrible. There is another OS kernel with a TPM driver now as well. We'd all better stop using this 'Linux' thing as well.
Indeed - the coffee room humour today was very dark :).
It is the leaders here in England that are justifying. The anti-war sentiment is pretty strong. Unfortunately the propaganda service which pretends to be the US media probably doesn't report this :).
My girlfriend is one of the girls pictured in the article - I concur.
And here, I'll be popping down later with my screwdriver set, breakout box and laptop to see if I can give them a reprogramming they'll never forget...
Since the two postscripts are different, changing the same letter in both will NOT result in the same hash.
Actually the textual part of both is the same, they just found some PS magic to choose between two 'sub-documents' in the same PS. Indeed one can trivially show that modifying both files will result in a collision since both files contain the text of both messages.
OTOH, in the real world, one would be hard pressed to change the contents of an existing document.
Problem: all of the new fat binaries will take up twice as much space on my already crowded hard drive.
/Applications/GarageBand.appt ions/GarageBand.app rjw57$ find . -type f -perm -0770 | xargs du -h -c 2>/dev/null | tail -n 1
But it wont really matter. Take GarageBand, a big heavy app. We'll only need to double the size of the executables:
persephone:/Applications/GarageBand.app rjw57$ pwd
persephone:/Applica
14M total
persephone:/Applications/GarageBand.app rjw57$ du -h -c 2>/dev/null | tail -n 1
63M total
So we go from 63Mb to (63+14)Mb = 77Mb. An increase of about 20%. A 2Gig OS X install goes to 2.4Gig. In these days of O(100Gig) HDs does it matter?