The Side Channel

« first day (2896 days earlier) ← previous day next day → last day (1783 days later) »

12:05 AM

@SqueamishOssifrage Removed it from HNQ

I have been reconsidering how to talk about inverting a hash function. I used to go to there not being a true inverse of the function because it is not injective. However I don't think that is helpful now. It's common knowledge that this isn't the case for hash functions because almost everyone has heard of hash collisions.

There are generalizations of the idea of inversion which apply to otherwise non-invertible functions. I'm not familiar with the terminology however.

When it comes crypto hashes, I'm looking at drawing a distinction between inversion meaning a preimage attack and inversion meaning inferring information about the input from the output. The latter including statistical artifacts which you might find when only considering a restricted subset of the domain.

... or almost equivalently, the second case could include me giving a very large set of messages for another person to choose from and asserting that I can compute which messages they chose if they give me hash(msg).

5 hours later…

forest

5:47 AM

@DannyNiu I strongly doubt each qubit is in full superposition with the others.

Even nearest-neighbor would be doubtful since Intel only has 50.

@Patriot MDC is useless anyway. However if you read the specs, they specify that they don't need strong security from the hash. Even MD5 would be fine in that situation. The problem with SHA-1 is collision attacks, which are not relevant in every situation.

@Patriot Even MD4 has not been "inverted", much less SHA-1.

2 hours later…

DannyNiu

7:42 AM

@forest That's a joke! The quantum computer runs the HOACS operating system. HOAX!! And its operating below absolute zero! I wrote that article just for the fun of it!

forest

8:10 AM

Guess this is twice I've been fooled by such jokes. :P

Patriot

@ DannyNiu Didn't fool me! So you are the author of the fake crypto paper at: web.archive.org/web/20190511120409/https://eprint.iacr.org/2019/… that Bruce Schneier talked about on his blog.

Patriot

8:32 AM

@ forest The most interesting word in that part of the specs on the MDC is "modest". The OpenPGP standard offers "modest" security, and that is good enough. (!) So, what in the world does "modest" mean when we are talking about SHA-1 and the MDC packet? However, even if it has been inverted, it is surely not an easy job to sort through the possible plaintexts. The inversion process would surely involve incredible computational power.

Do people use GnuPG with a "modest" expectation of privacy? The point here is that the puzzle palace folks would not invent something that does not have a crack in it when they use their super-tech insights. Not going to happen. The inversion of SHA-1 is a big prize, the one thing that most agree cannot happen, the single point on which to make a lot of encryption fail--and it curiously seems to stick around.

DannyNiu

9:06 AM

@Patriot I've posted paper to eprint, but they all get rejected or retracted. I live behind a national firewall so I cannot access web archive, could you hint me on the topic of the said paper?

DannyNiu

9:26 AM

Never mind, I found it

Anyway, I didn't wrote that paper.

Crypto HNQ

Q: Yarrow and weak cryptographic keys on macOS

Given that macOS uses Yarrow as a PRNG, and Yarrow is only capable of 160 bit outputs (/dev/random on macOS is actually Yarrow) are keys generated on macOS weaker than Linux? Even BSD has switched to Fortuna. Is this is big gaping hole in Modern macOS security? https://en.m.wikipedia.org/wiki...

Patriot

9:57 AM

@DannyNiu That paper was a quirky hoax, so I thought you might have written it.

4 hours later…

Patriot

1:52 PM

@ Ella I hope you realize that this website just disheartened a well-meaning contributor when that person's question was deleted--I refer to the recent question about English words in the one-time pad, What bothers me is the blatant disregard for the rules that allowed a recent question about the beauty of proofs to float, while then turning to the rules to get rid of a question in which the responder was rude to the questioner. I think this was poorly done.

This website should make room for questioners whose inquiries are very basic indeed. These people do in point of fact ask important questions once in a while.

Prejudice against questions involving the one-time pad should not stop us from acknowledging that the OTP can be a good first step for many to get involved in cryptography.

Ella Rose

I didn't delete anyones question. It appears that you are referring to this question, which is clearly still available on the site.

What I did was remove it from the Hot Network Questions list, per what the community decided in this meta thread

Similarly regarding the beautiful proofs question - I personally opposed it, but the community voted to keep it, and they are the ones that make these decisions.

Patriot

I see that. What I am telling you is that you have decided to follow the rules in case X but not in case Y.

Or the moderators were encouraged to do so.

Thank you for your attention and rapid response.

Ella Rose

Except that what you are claiming is simply not true? I did not delete the question you are going on about, so I did not follow the rules in case X but not in case Y. The question you are upset about has an accepted answer on it. So the website did not "dishearten a well-meaning contributor when that person's question was deleted" - because this literally did not happen

The user that asked the question is the only one that can click the accept-answer button. So clearly they were satisfied with the response, and not disheartened.

What exactly is the recourse you were hoping to achieve by bringing this up?

1 hour later…

Patriot

3:08 PM

Can we cancel this one, pursuant to the policy established on meta?
(cancel it as HNQ, not close/delete it)

Future Security
Future Security
7:55
Not my decision, but no one time pad question I can think of is imo worthy of HNQ status.

Ella Rose
Ella Rose
@SqueamishOssifrage Removed it from HNQ

I see, you did not delete it, you just removed it.

I think it is reasonable to assume that the questioner will be disheartened when he or she realizes the question was removed.

Removal happened after the questioner accepted the answer.

Removal certainly will have a chilling effect in some cases.

Ella Rose

> I think it is reasonable to assume that the questioner will be disheartened when he or she realizes the question was removed.

Not necessarily. They already have the answer they wanted. Being on HNQ would attract more answers, which is not necessary if the question is already answered.

> Removal happened after the questioner accepted the answer.

This is largely irrelevant. Considering the previous statements made above, this is completely harmless. If it was removed before an answer was accepted, then conceivably it could have affected/influenced that outcome. However, that is the exact…

Patriot

Rudeness towards questioners is something which should not happen.

Ella Rose

Finally, something we can agree on. However, if the questioner did not flag any content as rude/abusive/etc, then there is nothing to worry about.

The user said thanks for the answer, they were clearly not upset or offended by anything

Patriot

3:23 PM

That is an important point.

I too am very glad that the world will continue in its course. You dealt with my complaint very quickly.

fairly, thoroughly

4 hours later…

Squeamish Ossifrage

7:30 PM

@SqueamishOssifrage No, but what should I do? — R1w 2 mins ago

wat

(response to: Can you identify what you are substantively asking that is not already addressed in the other questions?)

4 hours later…

Crypto HNQ

11:35 PM

Q: Was the Enigma's double stepping mechanism intentional?

It's sometimes refered to as the double stepping anomaly, so was it just a design flaw or was it put in place deliberately?

encryption classical-cipher substitution-cipher enigma

forest

11:47 PM

@Patriot MDC is silly. The OpenPGP standard is not great. It uses non-standard CFB mode and lacks true symmetric integrity protection, such as that provided by an HMAC or AEAD cipher. It's a very complex and bloated format. However privacy is not the purpose of MDC, but (weak) integrity to protect from some very specific types of attacks. Use a digital signature if you want "real" integrity.

-1

Q: Blockchain help a software system more secure?

I think blockchain will make data in a software system consistent, resilient and preventing from malicious information. So we can state blockchain bring more secure to a software system. What do you think about this statement?