Won't the mp3-players be as useless as a betamax-player for the general public, as the copy-controlled cd's becomes more and more common?
sigh The copy protection that is being added to CD's is woefully inadequate to prevent it being ripped without serious backing from every CD drive maker on the planet (or alternatively every digital music player in the world...). The existing set of CD players have to be able to recognise the bitstream on the CD as music (i.e. sufficiently close to the redbook that they will output a decent enough audio signal). Therefore a sufficiently intelligent combination of hardware/software will also be able to do the same. And it doesn't even have to be that intelligent...They just have to recognise bad/conflicting data and assume/be told that it is to be interpreted as cd audio with the necessary interpolations over bad data. Copy protection requiring hardware agreement in a non closed (i.e. no paladim) world is unlikely to work - see the sites showing you how to hack out the copy protection on minidisk..the reason this is not a terriby big deal on mini disk is that the transfer medium is a physical so passsing around dupliates is expensive and time consuming compared to kazaa/ gnutella / lan parties / warez medium of choice. CD copy protection has a very short life span in my book because there is very little likelyhood that the DMCA could realisticly be applied to hardware/software that corrects 'damage' because that is all that can be done in the existing framework - damage the data.
Now if they comeup with a solution that replaces CD's with DRM built in it's a different proposition - still breakable perhaps depending on the algorithms and implementation (and of course on the legal system backing it up). But there's no way to bolt on DRM to something like CD's, and things like DVDA don't seem to provide much more to be honest. Your mp3 player will have a lifespan longer than its avaerage lifespan methinks.
you send a message of X bits, let X be three for argument. Lets assume you don't need to send a terminator identifier to simplify the proceedings.
0110100101100010 xxx
now we need to send the next OTP in the message this will be O
this leaves us with
0110100101100010 xxxooooooooooooo
so the sixe of the next OTP is 16 - 3 = 13.
Repeat a few times an you run out of pad. This process gains you no advantage over just having a big OTP and just moving through it.
Some people might say compress the new OTP so it fits in smaller remaining bits...
since the key is random (if it isn't you have a shit OTP) it cannot be guaranteed to always be compressible (in fact it almost certainly won't be) or, if you rebuilt the key repeatedly till you found one that was then your key has ceased to be random so your OTP is no longer an unbreakable OTP and just a crappy piece of bad crypto masquerading as one.
And you only ever use the pad once.
very important that.
A = plaintext 1
B = plaintext 2
X = onetime pad
A+X = A'
B+X = B'
an evesdropper knows A' and B' so
A' - B' = (A + X) - (B + X)
= A + B + X - X
= A + B
this still looks like gobbledegook but it is a very different sort of gook (or is it gobble).
Eve is now left with data formed by joining two non random pieces of data. Assumptions (educated guesses) about both A and B can be made which quickly yield info (and as anyone who's ever tried basic frequency analysis knows the moment you get a few bits from a guess the rest just starts falling into place).
Slashdot Mirror
The existing set of CD players have to be able to recognise the bitstream on the CD as music (i.e. sufficiently close to the redbook that they will output a decent enough audio signal). Therefore a sufficiently intelligent combination of hardware/software will also be able to do the same.
And it doesn't even have to be that intelligent...They just have to recognise bad/conflicting data and assume/be told that it is to be interpreted as cd audio with the necessary interpolations over bad data.
Copy protection requiring hardware agreement in a non closed (i.e. no paladim) world is unlikely to work - see the sites showing you how to hack out the copy protection on minidisk..the reason this is not a terriby big deal on mini disk is that the transfer medium is a physical so passsing around dupliates is expensive and time consuming compared to kazaa/ gnutella / lan parties / warez medium of choice.
CD copy protection has a very short life span in my book because there is very little likelyhood that the DMCA could realisticly be applied to hardware/software that corrects 'damage' because that is all that can be done in the existing framework - damage the data.
Now if they comeup with a solution that replaces CD's with DRM built in it's a different proposition - still breakable perhaps depending on the algorithms and implementation (and of course on the legal system backing it up). But there's no way to bolt on DRM to something like CD's, and things like DVDA don't seem to provide much more to be honest. Your mp3 player will have a lifespan longer than its avaerage lifespan methinks.
Fraid not sorry
.
you have an OTP pad of (say) 16 bits
0110100101100010
you send a message of X bits, let X be three for argument. Lets assume you don't need to send a terminator identifier to simplify the proceedings
0110100101100010
xxx
now we need to send the next OTP in the message this will be O
this leaves us with
0110100101100010
xxxooooooooooooo
so the sixe of the next OTP is 16 - 3 = 13.
Repeat a few times an you run out of pad.
This process gains you no advantage over just having a big OTP and just moving through it.
Some people might say compress the new OTP so it fits in smaller remaining bits...
since the key is random (if it isn't you have a shit OTP) it cannot be guaranteed to always be compressible (in fact it almost certainly won't be) or, if you rebuilt the key repeatedly till you found one that was then your key has ceased to be random so your OTP is no longer an unbreakable OTP and just a crappy piece of bad crypto masquerading as one.
And you only ever use the pad once.
very important that.
A = plaintext 1
B = plaintext 2
X = onetime pad
A+X = A'
B+X = B'
an evesdropper knows A' and B' so
A' - B' = (A + X) - (B + X)
= A + B + X - X
= A + B
this still looks like gobbledegook but it is a very different sort of gook (or is it gobble).
Eve is now left with data formed by joining two non random pieces of data. Assumptions (educated guesses) about both A and B can be made which quickly yield info (and as anyone who's ever tried basic frequency analysis knows the moment you get a few bits from a guess the rest just starts falling into place).