« first day (209 days earlier)      last day (4663 days later) » 

3:00 PM
@Jin Cheers!
@Iszi heh "Vista not well received by business" - understatement of the decade!
@Jin - did you see the couple of bug reports I popped into meta. It looks like they are IE7 specific, but also only happen with security.SE so probably something incompatible in the CSS
@RoryAlsop ok i'll look into it. although we officially no longer support IE7.
@Jin Does anybody?
i figured that was probably the case - it is no big deal - I can remember the URL's I want to go to, and it is only when at work as the client only gives me IE7
Except PaperThin, I suppose...
@ScottPack Hey, I still have clients on IE6. Actually, I still have one running NT
in some limited areas
3:03 PM
Well, they may have NT on site, but whether or not they run it.... :)
apparently "apps might break"
if upgraded
I know many big corps/gov still use IE7, or even IE6! but i figure if you're an IT Pro working in those environment, chances are you're free to install other browsers on your desktop
@Jin depends - we tend to argue that security folks should have more things locked down than others, as security people are a big risk (they're very clever and sneaky, you see:-)
and it goes similarly for IT bods
Now I have to google "IT bods" >_<
@RoryAlsop but aren't older browsers even less secure than new ones?
IE6 still has that thing where the site can grab the content of your clipboard without notifying you.
3:05 PM
@Jin I would argue that the issue you would run into is more about desktop management and less about what we want to run.
@ScottPack I agree.
@Scott has it absolutely right
See @Rory's comments about "apps breaking" and my veiled reference to PaperThin.
As I recall, they only recently started supporting IE8 for their CMS.
I would love it if we could just mass enforce all orgs to upgrade browsers to IE9, FF5 and latest Chrome, Opera and Safari - but they can't do it so they manage risk into little pockets. The risk of everyone being on IE6 is often easier for them to cope with than some folks being on other browsers that may or may not be outside simple management processes
Ooooh - I had an interesting glitch on that post @Jin - once it got really long lots of little arrows appeared all over it until I hit Upload:-)
very odd
@RoryAlsop Can you clarify? What sorts of security risks are commonly risked due to "management" outages for more recent browsers, that would trump the many ie6 bugs like grabbing passwords out of the clipboard? Is there a good Q/A in there?
3:18 PM
to be fair, I was glibly running two points together. I definitely don't approve of IE6/NT/anything that outdated and known vulnerable, however I also understand some of the risks around letting highly skilled individuals have versions which may not be properly supported/understood by support teams etc
What does it take to migrate a question from crypto?
Q: advances in usability for cryptography/authentication

Jason SI'm wondering if there have been any recent advances (say, the past 5-10 years) in human usability for cryptography and/or authentication? By that I mean something that makes it easier for an average person to make use of the benefits of cryptography. It strikes me that although we have all the...

@nealmcb Probably need one of the crypto diamonds to do it. Can't even see that unless you're in the private beta. (Unless all site diamonds get some special uber-access rights.)
I don't think there are any diamonds there yet - how would we tell?
Oops - sorry for posting a closed beta link....
@nealmcb There always are.
@nealmcb Check the bottom of the "about" page.
(Apparently, you can't even see that much if you're not in the private beta.)
"The following users are our current community moderators:

3:23 PM
@RoryAlsop which post?
Now we just have to create a User name TBD :P
@Mvy FTW!
Area 51 has only 5 "avid" users listed. Probably one or two of those know something about it.
I'd almost be surprised if @ThomasPornin didn't have some connection to (or place among) the diamonds there.
@Iszi I am not sure there are diamond-enhanced people in the private beta
Dunno why, but I find it really annoying that we can't even get read-only to private betas.
3:27 PM
Well that would negate the "private" thingy :P
If I have diamond powers then sure nobody told me !
@ThomasPornin - Have you gotten your rep capped today?
@Iszi Not yet. Currently 12 upvotes on security.SE and 14 on crypto.SE for today.
@ThomasPornin Slacker.
3:51 PM
hello sec.se, you look nice today! :-)
@RoryAlsop oh you mean a chat glitch
@StefanoPalazzo Thanks. We try :)
I'm not sure how I feel about the mass of blue names in here right now.
@StefanoPalazzo the new design is just a fragment of your imagination don't worry
"These are not the pixels you're looking for." /waveshand
Well time to run home :D
3:57 PM
Meh. I guess that's better than being told my kind isn't served here.
well, they look extremely nice anyway ;-)
@Mvy Same here. It's lunch.
hm... one hour and it's dinner here :P
So back in a few.
someone mentioned the "NSA algorithm stack" in a comment, is this the list they're talking about?
The vast majority of the National Security Agency's work on encryption is classified, but from time to time NSA participates in standards processes or otherwise publishes information about its cryptographic algorithms. The NSA has categorized encryption items into four product types, and algorithms into two suites. The following is a brief and incomplete summary of public knowledge about NSA algorithms and protocols. Type 1 Product :Main article: Type 1 encryption A Type 1 Product refers to an NSA endorsed classified or controlled cryptographic item for classified or sensitive U.S. gover...
I know it's on topic, but I thought I'd ask anyway :P
4:16 PM
@StefanoPalazzo Seems likely, but I haven't heard that exact term. What comment?
@nealmcb security.stackexchange.com/questions/5243/… the first one by Andrew Anderson
Now I wonder what stronger symmetric key algorithms are out there
and, in continuing that thought, why nobody uses Rabbit. It's all green in Wikipedia's comparison table
@StefanoPalazzo Rabbit is quite new (2003). Also it was patented until 2008 (and then pushed to public domain)
None of the eSTREAM portfolio stream ciphers are widely used
@ThomasPornin does that necessarily mean they have had less scrutiny?
@StefanoPalazzo Less than AES, certainly
@StefanoPalazzo Got it. Hmmm - those comments just disappeared - hope they're going into the answer....
4:26 PM
They still got a fair bit
@nealmcb yeah you missed it by a minute :)
Stream ciphers make sense only insofar as they offer better performance than AES
but AES performance is already quite adequate most of the time
hence, low usage
also, no actual published standard complete with bells and whistles (e.g. an allocated OID)
I'm getting close to 40MiB/s out of AES, that's much more than adequate. The key derivation takes much longer than encrypting any data set that fits into memory
so, I suppose that means that I really have no need for a real stream cipher
As I recall the reason for AES 256 was for protection against the possibilty of quantum computing attacks
So, AES doesn't rely on the difficulty of integer factorisation?
(yeah I know very little about crypto... but I want to do it properly anyway)
or why does the quantum computing scenario matter?
@ThomasPornin that's a great comment about the exponent size in DH - thanks a lot!
I was using 620 bit exponents until I came to sec.se :-)
4:34 PM
@StefanoPalazzo It is... a good question
the quantum computer is mythical
current record is a quantum computer with seven "qubits"
right, I understand that in cryptography, it's another way of saying "integer factorisation becomes easy" (that's just the feeling I get)
A quantum computer is a device for computation that makes direct use of quantum mechanical phenomena, such as superposition and entanglement, to perform operations on data. Quantum computers are different from traditional computers based on transistors. The basic principle behind quantum computation is that quantum properties can be used to represent data and perform operations on these data. A theoretical model is the quantum Turing machine, also known as the universal quantum computer. Although quantum computing is still in its infancy, experiments have been carried out in which quantum...
a QC should be able to break a symmetric cipher with a n-bit key in about 2^(n/2) steps
That article mentions AES and Grover's algorithm
But like Thomas says, don't sweat the quantum threat....
a QC will also factorize big integers efficiently
factorization breaks algorithms which rely on factorization being difficult (duh)
4:36 PM
> AES-256 would have the same security against an attack using Grover's algorithm that AES-128 has against classical brute-force search
and right now that's RSA, Rabin-Williams, Paillier...
@ThomasPornin How does it do for making coffee and folding space time?
Hmmm - actually I said that, and I kind of like the ring of it :)
@ThomasPornin also the diffie hellman scheme, yes?
@StefanoPalazzo Yes, there is a variant of the factorization algorithm which breaks discrete logarithm
including the elliptic-curve variants
nothing of the above really concerns symmetric crypto such as AES and SHA-256
4:38 PM
so basically, this applies
@StefanoPalazzo precisely
I actually got into security through learning about Off-the-Record messaging. Which has got to be the most beautiful security scheme there is
and it doesn't have that ↑ problem
@ScottPack Yes, it does coffee
so now, I'm trying to understand it properly by implementing it using another sort of crypto - which is why I'm posting questions about Skein and so on. lots of fun
(because the current prototype uses, as far as I know, quite a lot of power -- so you put a boiler on the transfo, and voila! hot water for coffee)
@StefanoPalazzo Implementing is a good way to learn. But remember that it is very difficult to know whether you implemented something secure (because a weak system compiles and runs tests just fine).
4:42 PM
yeah that's a bit frustrating sometimes. I'll never be able to use it
@Iszi Done.
@ThomasPornin Slightly less amusing than my presumption of the mythical "Solves All Problems (tm)", but much more practical.
actually now that I think about it, the NIST people should prefer an easy to understand algorithm over one with a larger security margin, shouldn't they? Must be reasonable to assume that most breaches are caused by implementation errors
like WEP and the Blackberry hack
For completeness, the old McElliece / Niederreiter cryptosystems (asymmetric encryption, signatures) currently appear to be immune to QC
@StefanoPalazzo Depends on whether they're using the algorithm themselves, or attacking its use by someone else :)
4:45 PM
@StefanoPalazzo The main job of a Spy agency in a modern state is to tell to the corporations of that state how not to get spied upon by foreign competitors.
@StefanoPalazzo You should link the actual comic instead of the image. That way we get alt-text. :-) Like so...
So when NSA says that AES is good, you can trust it, because US big companies will follow that advice and NSA certainly wants US big companies not to be hacked by non-US big companies
@Iszi nice, I didn't know we were oneboxing xkcd too :-)
There is a big history of Boeing and Airbus spying each other
(up to and including a Boeing-owned fake startup selling "encryption software" to Airbus)
4:48 PM
@StefanoPalazzo Someone edumacated me on that awhile ago. Love it.
so, there's an encryption plugin in the Pidgin instant messaging program. but, they way I see it, it's got a huge problem: It uses some type of public key crypto, which means if your private key is ever exposed, not only will that reveal your past conversation, it'll also authenticate it. So there's proof of the content and proof of the participants
but if you communicate completely without encryption, it's easy to forge. which means your participation in the conversation can't be proven
so the encryption actually makes you, in some ways, less secure
does that make any sense at all?
@StefanoPalazzo - Sort of, I suppose. Also, cryptography via IM client generally requires (depending on your usage, this may or may not be easily workable) the same client being used on both ends.
sorry, I'm all over the place today. Having far too much fun learning about cryptography
@StefanoPalazzo Strictly speaking, the private key is not used for signing the contents of the conversation.
It proves involvement in a conversation, but it says nothing about what was told
Also, it is possible to do Perfect Forward Secrecy (I don't know if the Pidgin plugin does it)
this means that the key exchange uses DH keys which were generated on the fly
it doesn't, that's why I'm now using off-the-record messaging, which does (using the socialist millionaire protocol)
4:55 PM
a transient DH private key, being recorded nowhere except in RAM (and only for a short time) should be immune to subsequent stealing
mh I don't understand this entirely. Let say I'm communicating with a rogue impostor
now I'm signing my messages (right?), that means that it can be proven, after my private key has been exposed, that the impostor was talking to me, no?
that is, without perfect forward secrecy (say using something like PGP)
@StefanoPalazzo It depends on the specifics, but a possible protocol is: at some point, an asymmetric key exchange takes place, and the parties authenticate to each other by signing the internal messages for that key exchange
afterwards it is just symmetric encryption and MAC
each party trusts the subsequent messages because of the MAC, but cannot prove to a third party that a given message is really from the peer
I am getting confusing, I need names
A and B run the key exchange and this results in a shared key K
call them alice, bob and evil eve, like wikipedia ;)
signatures have been computed and verified, both A and B are quite sure they talk to eachother
however, Bob cannot go to the police and say: here is a message from Alice, MACed with the key K, that's her message
because Bob also knows K
oh I see
that means bob can forge the transcript entirely
5:04 PM
and also the MAC can be verified only by knowing K, so if the police wants to verify the MAC they must learn K and be in position to fake the transcript themselves
the judge will find these proofs unconvincing
Now, if you have a man in the middle, can he participate in the key exchange so as to obtain the shared secret?
sort of like a three-way diffie-hellman, is that possible?
@StefanoPalazzo The signatures on the key exchange messages are meant to defeat that
I'm getting very confused now
if a man in the middle could obtain K, it wouldn't be a secure key exchange so that doesn't make any sense
@StefanoPalazzo yes
to prevent that situation, Alice signs her key exchange half (e.g. her transient DH public key if it is DH they use for that) and Bob accepts it only after verifying the signature
ditto in the other direction
that's extremely clever
you can do that exchange all the time and throw away the old keys to get perfect forward secrecy
5:10 PM
@StefanoPalazzo That's how SSL/TLS work with the "DHE" cipher suites
DHE meaning "Diffie-Hellman Ephemeral"
makes sense :-)
thanks for taking the time to explain this
@StefanoPalazzo @ThomasPornin is awesome like that.
Urgh. Maybe I'm taking this a little personally since there's a couple jabs and a down-vote aimed at me here, but I think this is getting old. Could we just close as NaRQ? I'd vote on it, but I lost that privilege about 24-ish hours ago.
Q: Protecting yourself from a malicious VPN.

RookI am connecting to a VPN server that has been compromised and is under the complete control of an attacker. The virtual network itself contains numerous attackers waiting for a new target. What steps should be taken to protect myself in this hostile environment? (Yes, I have to use a VPN that ...

@Iszi Of the chat active users, I think it's just mods, HB, and Thomas that kept it.
Again, I may be biased here, but it's comments like these that (I think) prove the case now: I fail to see how that is relevant. Its also shocking that this post got 9 up votes and its completely worthless. "Install an anti-virus hurrrr." and -1 this isn't an answer this is ridicule best suited as a comment. Both comments are to answers with 5+ up-votes.
@ScottPack Pity
And it doesn't look like we have any site diamonds on right now either.
was the last sentence in the question added after you've answered?
5:24 PM
@Iszi Although, that is expected behavior for the user in question.
@StefanoPalazzo - Nope. It was there before. And, frankly, I don't care. The only true measure that can be taken that offers guaranteed protection in that scenario is to not use the network at all.
The next-best measure is covered in paragraph 2 of my answer, though.
make sense, though that does explain the down-votes
@StefanoPalazzo Downvote singular. It's from the Asker. He has a bit of a...reputation, let's say :)
oh well, I wouldn't worry about it then
I'm surprised at how low his downvote number is
5:26 PM
people will falsely down-vote, that's just a fact of life on stack exchange
it's a good job an upvote so easily offsets that :)
@ScottPack What perhaps irks me most is that his rep is so high, and he seems to be not going (or staying) away, nor having any intention of changing his attitude.
Well, in his little sandbox he is good. The problem is that he just refuses to accept that there's a world outside, and it seems like his mission is to pat us on the head, tell us how cute it is that we haven't figured out the truth, and build schools until we convert :)
@ScottPack Agreed, and agreed.
Only, you know, by being an asshole about it.
5:30 PM
@ScottPack And therein is the problem.
I wonder what sort of answer he's looking for
@StefanoPalazzo At this rate, the kind that doesn't exist.
'Cause: 1.) His question apparently is lacking a lot of essential details which he doesn't seem keen to give. and 2.) For the question as written, there is no answer.
It's hard to tell what he wants from any of his questions. All too often the question he asks isn't the one he wants answered.
here's how this question looks to me: "What should I do if I need to visit the local mafia hangout, and they don't like me"
answer: send someone else
@StefanoPalazzo Translation: Use a VM or LiveCD. Both answers have been given, and rejected.
5:34 PM
maybe he's looking for the bullet-proof vest and jetpack type of answer, but that like you said doesn't exist
@StefanoPalazzo That's been given, too... and rejected, as well.
@StefanoPalazzo Again, it's hard to tell. I really get the feeling that whenever he posts a question he already knows what answer he wants, and gets really bitchy if he hears something else.
@ScottPack There's that, but I'm also beginning to wonder what hat he's really wearing.
yeah. well there's nothing you can do about that. I always tell people that we do this not for the people who ask the question, but for those who land on it via a google search. always answer (and indeed vote) with that in mind
if he disagrees, so be it.
Hell, half the time I'm doing it for me. I've learned more from writing answers than I ever did doing my job :)
I just get disappointed when people are properly disrespectful. I made allowances for humor and @Iszi.
5:39 PM
@ScottPack Heh... Thanks... I think.
do you guys get mad a very simple question on this site?
we kind of have them all the time at Ask Ubuntu, but I know some (like cstheory) get very mad at them
an example could be: "Is the salt stored in plain text?"
@StefanoPalazzo Refine that to "How/where should I store a salt?" and it might be more acceptable.
I was wondering a few days ago about the primes in DH
as in "Can I use the same, well known prime all the time in my DH implementation?"
@StefanoPalazzo I'd check with @ThomasPornin to see if that should fall under Cryptography.SE's jurisdiction, but it sounds fine to me.
5:46 PM
It can always get moved if need be.
they're in private beta still
Diamonds can't migrate to a private beta site?
pretty sure we can't
@ScottPack Unless they have permissions to the private beta, I don't expect they could.
@StefanoPalazzo A typical question which fits both security.SE and crypto.SE
In security.SE I would answer: "yes; make sure it is a good one such as those from RFC whatever"
In crypto.SE I would answer: "yes, because discrete logarithm hardness does not depend upon how the group is generated, and, indeed there is no secret data in the group" (followed by a discussion on elliptic curves)
5:53 PM
ono, I was going to answer my own question on that
but I can't for eight hours
this is so weird, I have at least 101 on every other site
Looks like a bad linkage
You should have gotten the bonus 100 rep
Q: Where do I get prime numbers for Diffie-Hellman? Can I use them twice?

Stefano PalazzoI realise it's very hard to generate suitable prime numbers and generators for the Diffie-Hellman key exchange. What is the best way to generate them? And if I have one, can I use it twice? According to Wikipedia, they are considered "public".

here's the question, if you want to go ahead. this is the answer I prepared in case it#s useful:
oho, it's gone (was in my clipboard) hehe
@ScottPack I lost 100 rep when I asked a question that was deleted as "spam"
I got an explanation, and something close to an apology, but the 100 rep were gone
I didn't know you lost rep for spam
So was the question undeleted?
but that's not so bad, it didn't fit the FAQ anyway
had they bothered to click on my name and take a look at my profile, they'd have realised it was an honest mistake. :) not that I'm bitter or anything :P
Speaking of assholes goes back to fighting with rpmbuild
5:58 PM
the question was a bit stupid, it was something like "where do I find viruses, I don't want to set up a honeypot, so is there a good place to get a bunch of them?"
Aw snap. That's an easy one
if $year <= 2000 then; goto USENET; else goto www.thepiratebay.com; fi
huh I didn't even think about that
indeed there are a few torrents like "collection of old computer viruses"
Curses! Canada attacks again! Took 3 tries before I realized why the vending machine was rejecting that quarter.
@ThomasPornin I'm going to take that as "document all of your newbie findings by playing jeopardy on this site" ;-)
@ScottPack A suboptimal strategy, since the Canadian quarter currently has higher value than the US quarter
Time to eat
so long, ladies and gentlemen
6:09 PM
@ThomasPornin I miss the good ole days when the USD was worth carrying sigh
Have a lovely evening, @Thomas!
6:49 PM
@Iszi, i don't like this color of hat thing much. Keep in mind that publishing information which is illegal in some jurisdiction may be ethical.
things like organizing demonstrations for democracy, videos showing soldiers killing civilians from an helicopter, it's hard to draw a line.
I just have to think at some childhood memories when visiting family members ones a year.
Crossing the border to east Germany: "Do you have news papers?", "No", "Open your bag". Pulls out shoes, pulls out newspapers from inside shoes. "What is this?"
Not, nice. Especially seeing your parents being afraid.
(In the end they somehow believed us that this was just a carelessness and we did not intend to smuggle news papers.)
7:21 PM
@HendrikBrummermann not quite the same, but: "do you have liquids?", "No", "open your bag". Pulls out toiletries, pulls out mouthwash... "Terrorist!"
cool, lotsa new faces here :)
@StefanoPalazzo problem with that "dillema", is that "security" is a vague fuzzy word.
Usually, it can be understood from context, but often that's not the case. That's why I find it best to avoid the word, unless its unambiguous by context.
The proper, full meaning of security, IMO, is really translated better into "risk management". Security is about tradeoffs, and finding the balance between conflicting elements of security (e.g. CIA, but also privacy aka confidentiality of identity), not just choosing to focus on only one.
What do you think about DNSSec?
7:42 PM
@thisjosh not being a networks guy, I think its a fantastic idea. That will never be used in large amounts (at least not in the next 8 years).
@AviD oh yeah. I'm still getting my head around it you can tell. Would it be fair to say your 'deniability is compromised'?
but, I'm not a networks guy, or a cryptographologisticianistic to know if the protocol is cryptographically secure... But, my point is its a smart move, and long past time, to get the base addressing protocol to be secure.
@StefanoPalazzo for shursies.
I'm all about the plausible deniability.... ;)
orly :-)
Then I can bother you with my beginner questions about that :P
I'm not sure what to think of it yet. My biggest concern is that it will deter people from using additional identification and authentication tools.
repudiation is definitely a security issue, but its usually the other way. On the other hand, since privacy is, to some extent, a confidentiality issue - but with a different risk perspective, this is from the pov of the user, and not the system - it comes back to finding the smart tradeoff.
much like you said - it seemed to be a bad tradeoff, in some situations.
@StefanoPalazzo go for it.
7:47 PM
I there any scheme that provides better, more plausible deniability than a plain text conversation?
actually I think @nealmcb might have some sharper insights on this - digital voting is (one of) his forte, and for voting the key elements are integrity+privacy.
@StefanoPalazzo but of course - double entendres.
I know for instance that OTR makes it easy to forge transcripts of conversations, but you'd still have to be a bit of an engineer to do it. Is there a scheme where it's plausible that even a non-technical person could forge a transcript?
I did not mean that!! I had no idea!!
yes, provable deniability, also refered to an an alibi
7:49 PM
@StefanoPalazzo in OTR specifically? or like in email?
in any kind of communication (i.e. I don't have a use case, I'm just very curious)
Between OTR, AES, PGP, and SSH, all my personal security needs are more than covered
also let me know if I misunderstand it. In my mind, deniability just means "it's easy to forge messages"
@StefanoPalazzo well then, emails are ridiculously easy to forge.
as are IM transcripts.
@StefanoPalazzo I think it would depend on the context.
might be "provably easy to forge", or "prove that someone else forged (or probably)", or even "prove that I could not have done it myself (ergo it must have been forged)"
I can also imagine a scheme where there is an incentive for random strangers to forge messages (specifically relating to your own communications), and in such a way that it's completely impossible to distinguish the fakes
btw @RoryAlsop congrats on beating me to 10K!! just noticed it....
we had a good race, but I flagged out between 8-9 k... ;)
@AviD I thought it was all about the Benjamins?
7:57 PM
@ScottPack no, the Charlies ;)
oo @ScottPack almost forgot, I found this for you today:
oh man whered i put it.... hang on
Should we start another room?

« first day (209 days earlier)      last day (4663 days later) »