The Side Channel

10:35 AM

@SEJPM congratulations, your voting answer is the accepted answer.

5 hours later…

3:05 PM

@kelalaka Welcome to the select club of trusted users on Cryptography! Congrats and thanks for all the effort you've put into our little community. Highly appreciated!

Only 25K site stats left in the privs, but if you need stats then just ask.

3:18 PM

@MaartenBodewes Thanks.

A quick question. I've cast a delete vote. Is there a delete queue that I could fail to see?

Great answer to pass the mark as well.

SHA-1?

Um, no, you're not missing anything. If you cast a delete vote then others will maybe see it, but as you can only vote once, you won't see a thing. Let me have a look.

No, the Blake one I think it was. But I could be wrong there :)

Hmm, I'm not even sure if there is a delete queue, I think they are just counted over the entire lifetime. I'll dig deeper.

I wish the authors were provided the answer. They, however, very clean about their design choices. I', appreciate their work. Lots to learn...

I missed my Cray machine, by the way :P

Buy one and put it... um...

3:24 PM

Well the power consumption...

Yeah, I can understand. I had some access to a 128 node AMD machine at university called the DAS.

There are small ones like Cray-1, Cray J-1 at the size of double refrigirator.

In a previous job they wanted me to put a server rack into my house.

Well I have seent hat a Cray can compute an infinite loop.

I happily declined.

3:26 PM

LOL.

Oh, I can also program an endless looooooooooooooooooooooooo

Actually, their compiler was so clear even those days that they eliminate empty codes. If you don't' use the result of the loop, it optimizes and eliminates the code.

Yeah, I've seem the same thing happening with the Java JIT.

Since then I always XOR results together and print them out afterwards. XOR is nice and fast and well, I'm a cryptographer :)

(not sure about the infinite loop though, Java is pretty good at guessing end-of-loop conditions)

guess?

Well, predict I suppose. It won't throw a dice, I would have noticed.

It sure buggered up my benchmark anyway.

3:32 PM

I know about the branch prediction of CPUs, however, newer heard about compiler or JIT

Dang, I almost made sense out of that comment :)

@MaartenBodewes If they paid for the electricity you would have had free heating.

You should :), waiting...

OTOH We're expecting a heat wave here in NL. Maybe you could leave the computer off and just use the water cooling though ;)

Server racks come with AC, right? :D

3:39 PM

Not these. Oh, Freon cooling, not water cooling.

Interestingly one of my most-upvoted answers is now "Just feed random stuff into SHA-1."

hot-network questions really skew the votes

Well, if those people found it interesting maybe they stick around. Votes always have been a popularity count, rather than a quality count. And it is certainly not an indication about the intricacy of a question / answer - possibly the reverse.

On the other hand, my next-most upvoted answer is basically "No."

A: Simulate int variables with byte or short

Although my X-mas special answer is now at 11 at least :) Finding out the best algorithms for doing multiplication, division & remainder using smaller word sizes, then allowing for signed numbers and then optimizing the heck out of it was ... a lot of work.

Christmas special answer. Fully tested using Unit tests, but as yet only on Java SE. Needs some work for instantiating the backing array. Some code can still be optimized by in-lining the left hand operand. Note that this code uses *= - assignment of the answer to the first variable - rather ...

so I guess I'm improving

3:49 PM

LOL

But negative answers are as good as positive ones, as long as they are supported by evidence.

A: Has AES-128 been fully broken?

Ah, but it's not!

No, AES-128 has not been broken by any means in any practical sense.

Um, I'm the one downvoter for that one :|

That's completely fair.

Well, at least you know the reason for it now.

There is something to be said about the brevity, I can completely get why people would upvote it given the question.

But more as an incentive to ask a better question.

4:50 PM

I think I'm the reason for both high votes. I wrote an additional answer to both than the questions to become HNQ.

It may or may not. Totally depending the day.

@MaartenBodewes I remember that answer, however, I was looking for another that you wrote a Cryptographic stack in Java. I would like to compare the code when converted to Kotlin. Do you remember any answer like that?

3 hours later…

7:35 PM

@kelalaka Sorry, was out. Um, a cryptographic stack? Sorry, I'm not sure what you mean by that.

1 hour later…

8:51 PM

@MaartenBodewes If I'm able to tell you I could found it :)
I cannot much define either. It was a great example of how to write a good crypto library, handling the errors, instance creation. Man, you had lots of ( good ) answers there. But that was special since it was complete.

A: How to implement password based hybrid encryption in java?

9:40 PM

It was similar to this one

You don't really want to generate the key pair from the password. The problem with that scheme is that there is no way to trust the public key. What is usually done is to encrypt the private key with a secret (symmetric) key that has been generated from a password. So, in addition to the hybrid ...

But not this one. It was ways better use of JAVA.

10:05 PM

@kelalaka Oof. I wrote too much stuff for SO, e complete cryptosystem right? I can try and find it, but honestly, I don't have much of an advantage over you in this regard.