« first day (3385 days earlier)      last day (1587 days later) » 

fgrieu
fgrieu
10:39
@MaartenBodewes The views are different (and could be slightly depressing for very different reasons) depending on if I'm looking at Stack Overflow logged, or in private view. I wonder if there's a third if I'm logged as an ordinary user.
Maarten Bodewes
Maarten Bodewes
10:54
@fgrieu I don't know, when I do this very few - if any - ads pop up, and I don't really see what the problem is to be honest. But maybe some kind of script is removing them for me. Strange...
fgrieu
fgrieu
11:07
@Maarten: When I access in private mode and scrool down I get a commercial offer to "Unlock siloed knowledge" starting $6 USD per teammate / month. It's not a third-party ad or a pop up, it's the meat of what the page is about.
Maarten Bodewes
Maarten Bodewes
I normally don't scrool around, but I'll have a look :P
@fgrieu Ah, that. I'm really not worried about it. The public Q/A could be a bit more pronounced, but I don't blame them for trying to make some money; in the end they need to get paid.
Seriously, if you want you can immediate click the Q/A block and be done with it. And most of the time you get there when searching through Google anyway. IMHO you only look at the standard SO homepage (with the "random" questions) if you're extremely masochistic.
They should probably replace it with a search bar, so you can skip that part. Would be less painful to the servers as well.
 
4 hours later…
Squeamish Ossifrage
Squeamish Ossifrage
15:33
0
Q: How does the entropy of $r$ influence the security of the one-time pad $F_k(r) \oplus m$?

X.G.In an one-time pad scheme, $s \oplus m$ is uniformly random for any $m \in \{ 0,1 \}^\ell$ if $s$ is uniform in $\{ 0,1 \}^\ell$. By the security of PRF, it seems to be secure to replace the truly uniform string $s$ with the output $F_k(r)$ of a PRF $F: \{ 0,1 \}^\kappa \times \{ 0,1 \}^\ell \rig...

one-time-pad pseudo-random-function entropy
The only things that are relevant here are the probability of collision between two values of r and the PRF-advantage against F_k. First model F_k as a uniform random function f; in this idealization the IND-CPA advantage against f(r) + m is bounded by the probability Pr[C] of a collision between two values of r, for otherwise you have a one-time pad. Then add the PRF-advantage against F_k, and the IND-CPA advantage is bounded by: Adv^PRF_f + Pr[C].
Squeamish Ossifrage
Squeamish Ossifrage
15:58
1
Q: Is there a universal construction for Davies-Meyer hash functions?

Keshav SrinivasanMy understanding is that there exists no strong pseudorandom permutation for which the Davies-Meyer construction is known to yield a provably collision-resistant hash function. Were I mean provably in the sense of invoking no assumptions like the ideal cipher model. But my question is, does ther...

hash block-cipher collision-resistance pseudo-random-function pseudo-random-permutation
The premise is irrelevant. Davies–Meyer gives the adversary control over the key, so everything about (strong) pseudorandom permutations instantly flies out the window.
Squeamish Ossifrage
Squeamish Ossifrage
16:50
@fgrieu In crypto.stackexchange.com/a/85633, maybe you should specify what hash function you mean with HMAC? For example, HMAC over Davies–Meyer with AES won't do much to avoid the timing side channels you mention!
 
3 hours later…
fgrieu
fgrieu
20:06
@SqueamishOssifrage Very right, that's done, thanks!

« first day (3385 days earlier)      last day (1587 days later) »