Yes, it does. It means you get features, bug fixes and new hardware support as they get developed, rather than waiting for rollup releases every six months or so.
Since ELF is the native binary format on linux, you don't really have much choice in the matter. You could compile it out if you wanted, but you'd never make it to/sbin/init.
the FSF did the same thing to Apple back in the early 90s over the look-and-feel lawsuits.
background reading on the subject can be found here, and here, and pretty much everywhere else google knows about.
if this kind of "boycott" did any good, slashdot would be using PNGs by now, wouldn't they? *cough* in this case, it's only a good idea if the linux community can come up with absolute, solid, will-hold-up-in-court proof that SCO is talking out of their collective asses about this.
and even then, who cares? SCO's unix products blow chunks anyway. SCO is irrelevant. they're being petty. why descend to their level?
probably a number that's irrational in a number system with a rational base will continue to be irrational in all number systems with rational bases. (this is in fact by definition; "irrational" means "not producable by dividing one integer by another", and bases are just notational division.)
if you really care to think about irrational number bases (and the associated notational headaches), then assuredly there is some number base wherein pi is rational (an "integer" ratio of the base). which isn't terribly useful information, unless you have a way of specifying irrational numbers to infinite precision, a way of notating them, and the mental flexibility to think in them. we wish you luck.
the more interesting question in terms of crypto is: if you know the value of the nth bit of pi, but not n (ie, you know the number but not the place), can you determine the value at the (n+1)th bit?
if you think about it for a second (informally of course) you'll see that you can't. in a true random stream the odds of any given number following any other given number are evenly distributed, which in binary means, even if you know if bit n is 0, the odds of bit n+1 being 0 are 50%.
therefore, you can use the bits of pi as a stream cipher. pick a starting offset (bit n of pi), calculate the successive bits of pi equal to the length of the message, xor it with the message, and transmit. the starting offset then becomes the key to the message, and if you transmit that securely (always a big if) then the protection is as good as a one time pad.
this works as long as either a) pi is not cyclic or b) the cycle (if any) of pi is longer than the length of the message. the ability to calculate the n'th digit of pi does not necessarily imply that pi is cyclic, although it might.
personally, i think this is hella cool. i might just have to write a/dev/pirandom driver; seed yourself at boot time from the {u,}random entropy pool, use that as the initial offset. it'd take some doing, bignum arithmetic and all that, but definitely doable.
Slashdot Mirror
Yes, it does. It means you get features, bug fixes and new hardware support as they get developed, rather than waiting for rollup releases every six months or so.
Since ELF is the native binary format on linux, you don't really have much choice in the matter. You could compile it out if you wanted, but you'd never make it to /sbin/init.
Good try though.
That joke pretty much writes itself, doesn't it?
except the Dead suck.
The Xorg Release Plan has a good overview.
i didn't know PHP was ever not an inappropriate language.
What's better than an uncompressed format for this sort of archival work?
FLAC.
IYTM "better than a lossless format".
Why did the $EUROPEAN_AIRLINE flight crash?
All the poles moved to the right half of the plane.
(no, i will not explain it.)
background reading on the subject can be found here, and here, and pretty much everywhere else google knows about.
if this kind of "boycott" did any good, slashdot would be using PNGs by now, wouldn't they? *cough* in this case, it's only a good idea if the linux community can come up with absolute, solid, will-hold-up-in-court proof that SCO is talking out of their collective asses about this.
and even then, who cares? SCO's unix products blow chunks anyway. SCO is irrelevant. they're being petty. why descend to their level?
i'd hazard a guess that pi, in base pi, is 1.
probably a number that's irrational in a number system with a rational base will continue to be irrational in all number systems with rational bases. (this is in fact by definition; "irrational" means "not producable by dividing one integer by another", and bases are just notational division.)
if you really care to think about irrational number bases (and the associated notational headaches), then assuredly there is some number base wherein pi is rational (an "integer" ratio of the base). which isn't terribly useful information, unless you have a way of specifying irrational numbers to infinite precision, a way of notating them, and the mental flexibility to think in them. we wish you luck.
the more interesting question in terms of crypto is: if you know the value of the nth bit of pi, but not n (ie, you know the number but not the place), can you determine the value at the (n+1)th bit? if you think about it for a second (informally of course) you'll see that you can't. in a true random stream the odds of any given number following any other given number are evenly distributed, which in binary means, even if you know if bit n is 0, the odds of bit n+1 being 0 are 50%. therefore, you can use the bits of pi as a stream cipher. pick a starting offset (bit n of pi), calculate the successive bits of pi equal to the length of the message, xor it with the message, and transmit. the starting offset then becomes the key to the message, and if you transmit that securely (always a big if) then the protection is as good as a one time pad. this works as long as either a) pi is not cyclic or b) the cycle (if any) of pi is longer than the length of the message. the ability to calculate the n'th digit of pi does not necessarily imply that pi is cyclic, although it might. personally, i think this is hella cool. i might just have to write a /dev/pirandom driver; seed yourself at boot time from the {u,}random entropy pool, use that as the initial offset. it'd take some doing, bignum arithmetic and all that, but definitely doable.