3,815 reputation
1625
bio website
location Wellington, New Zealand
age
visits member for 2 years, 7 months
seen 2 hours ago

I am an undergraduate computer science and mathematics student in New Zealand. My fields of interest are computer graphics, in particular the physics of light transport, and to some extent cryptography, as well as programming and software development in general.


Dec
22
comment Is a Mersenne-twister cryptographically secure if I truncate the output?
It's regrettable that the question as-is seems to imply that it's OK to do whatever the OP suggested. A shame for future visitors who may not have the insight of actually checking the rest of the thread, and end up using a homebrewed, brittle, and perhaps profoundly insecure PRNG as though it were a CSPRNG. I also like that you dismiss the entire field of cryptanalysis as if you were too good for it and just assume all your arguments are correct. Seriously, it's cool to be creative for experiments, but for a real project, no less an ONLINE CASINO GAME, you WANT to use best practices. NOT THIS.
Dec
22
awarded  Informed
Dec
21
comment How can uniformity of hash functions (e.g. SHA-256) be proved?
I'll leave it to someone else to write up a formal answer but I believe the short answer is that there is no proof. One can sort of offer mathematical arguments that SHA-256 is secure against such and such attack, or has such and such general property, but by and large it's just a matter of cryptographers around the world banging as hard as they can on it and seeing it if breaks. Symmetric cryptography is nothing like number-theoretic crypto which is much bigger on actual proofs (cf. factorization, DLP, DDH, ..)
Dec
21
comment Calculate entropy of key derived from PBKF2 function
@deltaaruna The salt is not (intended to be) secret, so it has zero entropy. The password has entropy $e$, and you use $2^k$ rounds - so the total amount of "work" in the computational sense to brute force the password is on the order of $2^k 2^e = 2^{k + e}$, so the "entropy" is $k + e$ (again, in the computational sense - the information-theoretical entropy of the key is at most the entropy of its inputs, but since it takes more work to verify your guess you can think of it as "extra passwords to check", e.g. 2 iterations = twice as many passwords to check computationally = +1 bit entropy).
Dec
20
comment Calculate entropy of key derived from PBKF2 function
(up to a maximum of the output length specified to the PBKDF2 algorithm)
Nov
30
comment Is this encryption scheme perfectly secure?
It's not really "the Caesar cipher", sure the Caesar cipher uses this primitive internally but is certainly not perfectly secure, so this is somewhat misleading. Otherwise good answer +1
Nov
26
comment password generation: random length or max length
@K.G. Quite true, and worth pointing out, though the increase in security is still negligible.
Nov
26
revised password generation: random length or max length
"smaller" to "shorter", 4 gramer!
Nov
26
revised password generation: random length or max length
fixed typo in equation
Nov
26
answered password generation: random length or max length
Nov
21
comment How does the length of the plaintext affect the cipher strength?
@RickyDemer I am wondering, why do you always use LaTeX code to format your comments? They tend to display rather oddly for me.
Nov
16
comment Are there any bijective one-way functions not based on number-theoretic hardness assumptions?
@SDL "hard to compute" has a specific meaning in cryptography (and as you probably guessed, it happens to mean "infeasible"). Maybe you should detail exactly what your requirements are in your question in unambiguous terms.
Nov
2
comment Is there an algorithm or hardware that can sign/verify natural time?
I guess the assumption was that time remained linear. Obviously if you can manipulate time through the laws of physics all bets are off... damn physics, always telling us "yes, but..." :p
Nov
1
comment Does ssl_rsa_with_rc4_128_md5 have known weaknesses?
@hunter Right.$\left . \right .$
Oct
31
comment Does ssl_rsa_with_rc4_128_md5 have known weaknesses?
It's unclear what you're asking. Secure with respect to what? Besides, the question is rather vague, and invites speculation - simply "using" a good cipher doesn't mean you are instantly and utterly protected, just like purchasing a bullet-proof vest won't protect you unless you wear it correctly.
Oct
21
comment Estimating random number entropy for input into 256 bit hash
@liquin Yes, that is correct, and, yes, that is correct. You would reject values 8 and 9 and try a new number until you get a value between 0 and 7 (which you interpret as three bits). The problem is the algorithm only works well if you have a large number of such numbers (preferably infinitely many) since you are probabilistically rejecting some of them. And to be fair using SHA-256 is probably the safest option, unless you have some requirements you have not told us about.
Oct
21
comment Estimating random number entropy for input into 256 bit hash
The last three bits are NOT uniformly distributed. 000 and 001 appear twice whereas the others appear only once. If you use these directly to build up a 256-bit bitstring you have a weakness. Either feed them into SHA-256, assuming around (but not exactly) 3 bits of entropy per number (you can calculate the exact entropy with Shannon's entropy equation) and collect the final digest when you've reached 256 bits of input entropy, or use rejection sampling to correctly extract uniformly distributed bits from a stream of such 0-9 numbers, assuming you have enough of them to avoid running out.
Oct
18
comment Selective format-compliant JPEG encryption?
You can encode data in the alpha channel for most photographs which don't use that channel (if the format permits it), which lets you introduce some redundancy, maybe saving the scheme. You can also insert encrypted metadata in most formats without breaking compliance, which may or may not be compressed. So I'm not too satisfied with this answer. There is more to it than just "you can't do it because there's not enough space". There is plenty of space in many image file formats to write a bunch of stuff (though obviously it will be very file-format-centric, not a general solution).
Oct
17
comment How exactly was the finalist chosen in the NIST AES competition?
Obviously the AES-NI instructions did not exist before AES was standardized, though it certainly helps performance a lot.
Oct
17
comment How exactly was the finalist chosen in the NIST AES competition?
I hope by implementation difficulty they really mean ease of implementation....