« first day (1888 days earlier)      last day (3290 days later) » 

01:11
man. bootstrap is a waste of time and effort
responsive design is not hard...
why throw in 300KB of CSS and Javascript when I can just do this myself much quicker :/
 
6 hours later…
07:18
@RоryMcCune dont remember who it was, but somebody tweeted "Just saw Van Wilder 3, and it was great!"
also, I knew she was cool, but never knew she was such a pottymouth.
07:45
Is it amateur day already?
@AviD Can you show some source that shows 3DES is not secure? Or not better then DES? (For compatibility reasons it is rather important for us to use 3DES instead of AES) — user200312 1 min ago
some pass me a mallet please
right??
@MarkBuffalo but why not? KBs are cheap!
@MarkBuffalo hahahaha just saw your message....
my condolences
@AviD Can you show some source that shows 3DES is not secure? Or not better then DES? (For compatibility reasons it is rather important for us to use 3DES instead of AES) — user200312 5 mins ago
@TildalWave I am tempted to close it as clueless
welcome to fkin 40 years ago
actually though it is not a SO question, but maybe a crypto.se
07:51
the question is where will he/she be a subject of ridicule the most ... then migrate there
this is a class example
dev logic: it works, therefore it is secure
I fkin hate devs, and I am one (kinda)
@TildalWave then here
Wait? @TildalWave didn't you used to be a mod? Or am I confusing you?
Crypto.se will approach it as a theoretical question
SO will tell him it's okay
actually, SO would probably tell him not to bother with 3DES, plain DES is just fine
@Arperum had to quit for mental health reasons :)
@Arperum @TildalWave oh! you lost your blue!
@TildalWave heard that
can totally understand
07:53
@TildalWave Ah, so I was not misremembering.
No no no, do NOT do this yourself. Cryptography is delicate, there is many maths involved. If you don't know what you're doing, you will mess it up. And, if you are trying to implement 3DES out of DES in 2016, you definitely don't know what you're doing. — AviD ♦ 1 min ago
@AviD Can you show official source which says 3DES should not be used in 2016??? — user200312 1 min ago
@TildalWave can you give him a source for that? The wikipedia page is ambiguous.
@AviD soz, I stopped being interested with "it works, I'm fine with that"
yeah, he basically already made up his mind.
saw my base64+reverse comment?
I did now
wanna bet he's gonna use fixed keys?
"it works"
4
A: 3DES with fixed-encoded random key

PolynomialThe security of this scheme largely depends on how good your key generation is. If an attacker observes two key exchanges, he may calculate the xor of the two unobfuscated keys: The values ka and kb are two unobfuscated keys. kA and kB are the associated obfuscated keys. o is the static obfuscat...

good grief I hate children
@kalina set fire to them
08:33
@AviD " there is many maths involved "
@kalina so do I, especially the kind that pretend to be adults.
@AviD what are you trying to say
@RoryAlsop I don't have the energy to invoke fire, I got woken up by being JUMPED ON
@kalina magik faerie dust on the tip of a unicorn's horn
@kalina isnt that a good thing?
@kalina even when it's your own kids, that's not great
@AviD lol
@AviD context :(
08:34
@RoryAlsop I like when they wake me up that way
@AviD jumping, not so good. Cuddling, much better
its when they wake me up by kicking me in the face, or sitting on my head, that I dont like as much.
actually - depends on how heavy they are
@RоryMcCune Are session keys necessary, or using fixed keys is also OK? with 3DES — user200312 7 mins ago
boom!
knew it
@RoryAlsop oh absoultley. I'm talking about the small ones.
08:36
@AviD should we not migrate to Crypto? ;-P
@RoryAlsop we should
he wants to know how to implement 3DES
done!!!
(heh)
@TildalWave well he does
problem is not crypto it's the scheme
08:37
@TildalWave pls that's the kind of thing crypto like talking about
it's not "is this secure" it's "I want to implement 3DES myself, how do I do that"
@RoryAlsop I was considering
@RоryMcCune if you have to ask...
but @TildalWave said it should go where he will get mocked the most, I think we do that best.
exactly
@AviD you don't think there are crypto mockers?
08:39
@TildalWave oh I agree absolutely, but now several people have said that and he appears undettered
so he might as well get some advice on how to make less of a mess of it than he would otherwise
@AviD if we're lucky Tom Leek will answer
@RoryAlsop oh sure. But they would answer the theoretical question of "how to do this", and not "I want to actually build a system in 2016"
@RоryMcCune ooh yeah
ping him
@AviD are you suggesting our crypto bretheren aren't concerned with practicalities
or maybe DW
@RоryMcCune not at all
dammit chris!
@user200312 3DES is still pretty secure: the best known cryptanalysis (on keying schedule 1 or 2) takes about 2^112 operations to crack. And 3DES is what is used by the credit industry in transaction requests by EMV cards, so take that as you will. — Chris 36 mins ago
he has a different opinion to you and called you out on it in public
you should nuke him
08:44
pffft
I'm serious, your entire reputation hangs in the balance
obviously the balance is counterbalanced by air, such is your reputation
@AviD Time for another anti OSI crusade:
7
Q: What do TCP/UDP add to "raw ip"?

John DorianI know that both TCP and UDP are built on top of IP, and I know the differences between TCP & UDP, but I'm confused about what exactly "raw ip" is. Would it be fair to say that TCP & UDP both implement IP, but that IP in and of itself isn't capable of transferring data? Or is IP some very low lev...

@kalina aww sweet
@CodesInChaos heh nah, I'm tired of that one
great news @AviD your favourite language just got a new release!
@CodesInChaos also JEEZUS on network engineering???
@RоryMcCune what's that? some kind of hip video streaming website?
08:46
In the ANSWERS
@RоryMcCune PHP 6?
@kalina pls, @AviD is waaay too hipster for PHP, thus his love of a combination of Kotlin and 4GLs
@user200312 Please see 3DES with fixed-encoded random key. You might be "OK with it" because "it works", but your users might not be and they deserve someone that knows what they're doing. I just hope I'm not one of them. — TildalWave 17 secs ago
@CodesInChaos actually only one answer, and not the top one. The top one mentions it for additional reading, for understanding the concept of layers - WHICH IS EXACTLY THE POINT
@RоryMcCune I wish I knew what you were talking about
wait a minute, no I don't
08:53
@kalina (words. hipster stuff.)
@RoryAlsop only you could call 4GLs hipster :op
@RоryMcCune lol
that look on your face when somebody sits down next to you with a surface 3 or 2
and you're just like =D
take a look at the new touch cover
look at this screen
"Though used earlier in papers and discussions, the term 4GL was first used formally by James Martin in his 1982 book Applications Development Without Programmers"
poor you
08:54
@kalina SP4-Master-Race!
oh you have a surface 2! that's so adorable, look how tiny it is
@kalina hahahaha - sizeist
oh don't mind me, I'm just playing civ 5 on my surface pro four.
"it's the motion in the ocean" vs "who really wants to travel to hawaii on a row boat?"
@kalina Thor Heyerdahl?
bla bla bla and some people did the smaller end of the scale as well
that doesn't make it right
08:57
lol
I know
also, LYNE is amazing for passing hours of my life
@kalina OK, but can it make toast?
@TildalWave who needs a tablet that makes toast
I do :P
it keeps my hands toasty if I hold it right
top right corner
08:59
@kalina All hail Ghandi
@Lighty pls ghandi has double unhappiness from # of cities
he's great if you're going for a four city cultural victory but once I get to the atomic era I want to nuke everybody
Google Chrome is such a power whore as well
I had to uninstall it
it was responsible for 40% of my battery usage during sleep
@kalina Was referring to a very old bug/meme/whatever, but ok, I play as the dutch anyways because go netherlands
Now I'm curious, what does @kalina think of that new Civ game in space? :P
@Lighty pls rome because faster building production if building exists in capital
@kalina ooo look who's complaining about Chrome hogging up your cool
@Lighty not as good as civ 5
@TildalWave pls I use edge
09:02
IIRC you had things to say when I moved to Firefox from Chrome for similar reasons
I have things to say about everything
@kalina finally something we can agree on, I like it, but it feels so watered down, since Civ 5 felt much larger and better xd
@AviD Two answers talk about Layer 3/4. Numbering layers makes no sense.
@Lighty what do you mean finally
09:04
@CodesInChaos exactly! Literally no actual real world apps use the OSI model.
They are all roughly around about a bit like parts of OSI
@kalina shrugs If you literally translate "Finally something we can agree on" to dutch, then it becomes a Metaphor like thing over here, so I thought you English speaking beings had something similar to it too
I dislike OSI even as a concept.
"Literally no actual real world apps use the open source initiative model"
is that what that sentence says?
Anything above IP can be replaced, split or reordered at will.
@kalina Open Systems Interconnection
09:06
The Open Systems Interconnection model (OSI model) is a conceptual model that characterizes and standardizes the communication functions of a telecommunication or computing system without regard to their underlying internal structure and technology. Its goal is the interoperability of diverse communication systems with standard protocols. The model partitions a communication system into abstraction layers. The original version of the model defined seven layers. A layer serves the layer above it and is served by the layer below it. For example, a layer that provides error-free communications across...
@CodesInChaos yup
don't make me fulfil @Simon's role of calling you all nurds
@kalina don't fill @Simon's anything pls
@CodesInChaos sure it does, you can number it sequentially.
well I'm sure @Simon would do the filling
he would be right up here calling you all nurds
is still early you can't talk so much nurd
09:07
@AviD But for that to make sense you need a fixed ordering.
up until TCP it makes sense, and is almost-but-not-quite de facto standard. Above that it makes sense as a relative notation, not absolute.
sense:absolute;
@AviD IMO even TCP doesn't make sense anymore.
right so that's four, just three more
TCP does four things at the same time.
09:09
@CodesInChaos oh?
@CodesInChaos okay...?
1) The port field, so you can have several services on the same machine. 2) It forms a connection 3) It does flow control 4) It does reliability
sure. so?
its not different layers?
IMO all of these deserve a separate layer that can be swapped individually
and 2-3 mean it clearly includes higher layers than UDP, so putting it at the same layer doesn't make much sense.
@CodesInChaos 1 and 2 need to go together. 3 and 4, while seperate, still make sense together. Moreover, putting 3 and 4 as aspects of the connection (#2) makes sense too.
@CodesInChaos well thats the difference between TCP and UDP.
I actually had to explain that difference this week, to my high school class.
I did have to simplify it a lot ... :-(
the spirit of Tim BernersLee is turning over in its grave.
YES I KNOW
But if, for example I want to replace the connection level, to add connection level crypto, fusing all those layers in one protocol prevents that.
Or I want to swap the flow control for LEDBAT or whatever fits my application.
09:14
@CodesInChaos huh dunno wat ledbat
@CodesInChaos I dont see it.
@AviD flow control that torrent uses (on top of UDP) to reduce disturbances of other connections.
the way I see it, you have 3 choices: 1) you want the connection, and everything that implies; 2) you DON'T want the connection, and everything THAT implies. 3) you want to do something custom and non-standard.
@CodesInChaos yeah reading
@AviD For example I don't like the connection feature of TCP because it sucks (No crypto, vulnerable to RST attacks, etc). So I swap in another protocol that's better, but reuse the flow control and reliability layer of TCP.
> and is used by Apple for software updates
@CodesInChaos but what you are saying would lead to splintering. IMO there is much value in having a few standard choices.
the standard protocol sucks? improve it.
dont split it out to 13 similar, yet ultimately different, protocols.
Bernstein's CurveCP comes closest to what I want, but it suffers from the typical Bernsteinisms.
09:18
@CodesInChaos whooosh
heh your blog is #5 for that term.
It's UDP->Crypto+Connenction->Flow Control->Reliability
but but but! UDP is connectionLESS
@AviD You build the connection on top of UDP.
@CodesInChaos what are typical Bernsteinisms?
@CodesInChaos again, why dont you just fix TCP?
@AviD Increasing the modularity of TCP is the most important fix TCP needs.
09:22
hmm good point.
@AviD Half finished code. Extremely unusual code structure. A rule the world philosophy.
kinda like how SSL as a protocol supports plugging in different ciphers.
@CodesInChaos one key to rule them all?
I'd like flow control layers and reliability layers as simple libraries that can be used from user-mode.
@CodesInChaos oh no that I would NOT like at all.
because programmers are gonna go and mess up the hell out of that.
09:24
@TildalWave No, it expects to have one crypto-termination service at the OS level. And then expects applications to communicate with that via pipes/file-descriptors.
at least now, there is very little they can do wrong at those layers, and they confine their fuckery to the higher levels.
@CodesInChaos not sure that's not even worse :O
@CodesInChaos so he's brilliant at pure mathematical crypto, and sucks at the real world implementation?
@AviD The implementation is also brilliant. But only if you have the freedom to reinvent your whole software and OS stack at the same time.
And you buy into his rather unconventional style.
@CodesInChaos hmm, maybe a better word for what I meant would have been deployment.
as in, not the code writing, but the using.
UX and etc.
09:28
Yeah. He sucks at finding a path from the current messy world to what he imagines as ideal world.
the problem is, even in an ideal world, you still have messy humans to deal with.
is IIS susceptible to Heartbleed?
and messy applications written by messy programmers.
@kalina no
@kalina No. It uses MS's crypto lib (SChannel), not OpenSSL.
heartbleed is OpenSSL only. IIS does not use that.
unless your app uses it, e.g. in PHP.
09:30
Or you use a reverse proxy/SSL terminator in front of IIS.
you know quick answers like that is basically the reason I idle in here
@CodesInChaos sure, but then its not on IIS...
@kalina heh
Some of Bernstein's colleagues build an OS (EthOS) that fits with DJB's ideal model.
I always assumed it was the simonmocking and roryworship.
@CodesInChaos I'm sure the UX is fantastic.
@AviD :-)
10:08
@RoryAlsop how do you ping a site's mods in TL?
that DES-ECB bloke asked on SO...
-10
Q: Implementing 3DES using DES

user200312I have existing class which performs the DES computation. Can I implement 3DES from this class, if yes how? Also what are best practice to use 3DES? Keys, encryption mode, etc. PS. Is it possible to use 3DES for MAC? PPS. Currently I have such algorithm for MAC HMAC := DES(block1 xor block2 ...

actually getting decent responses, at least in the comments.
I understand but current one works like this and I think it will give some "reasonable" security, I don't want security against NASA — user200312 3 hours ago
Amazing. I think NASA has other things to do than hack stuff, they'll leave that to the NSA.
@AviD [site] mods: [message]
@TildalWave ah cheers
10:26
The moment you use XKCD comics in confidential papers to explain stuff and get bonus points from boss for being clear with explanation
@AviD 3DES is still "approved" by NIST as long as it is the three-key variant (see NIST SP 800-131A rev 1).
@ThomasPornin yeah I know
but he wants to use the 2key variant
with fixed keys
oh yeah, and ECB
@AviD But that's not the worst part of his custom MAC scheme.
0
A: Implement 3des using des

Thomas PorninCalling "HMAC" that which is not (at all) HMAC can only lead to confusion. Though DES is, by today's standards, too weak for comfortable usage, this is not the weakest part of your custom MAC. In fact trouble begins right at the XORing of the blocks together. For instance, the attacker may swap ...

and still and besides, even so, I would not consider 3DES to be a wise choice, even when used properly.
@ThomasPornin oh yeah didnt even get to his mac
woot a bear answer!
and one I totally understand too.. about crypto :-O
> This is a bit like driving a 40-ton truck: being extra careful is not sufficient not to run into trouble; you really need to first learn how to drive.
and another classic bearism.
@ThomasPornin (in your notations, that K1 and K3 may be distinct from each other) -> surely you mean "MUST" be distinct?
and yet and even so, telling him to implement new code with 2016 is still not a good idea.
you're right though that its not the worst part of his scheme, but still.
@AviD No, it is still a "may". In the three-key variant, the three keys are supposed to be chosen uniformly and independently of each other, which means that they are still allowed to be equal out of pure chance.
10:35
okay, then at least should.
actually you said "distinct", not "unequal". So semantically, you do mean "must".
If you say "MUST be distinct" then people will begin to do something weird like actually checking that they differ, or force some bits to specific value to ensure that they are distinct.
the more we engage with this user, the more I think all these answers are wrong: the correct answer is simply "step away from the keyboard before you hurt somebody".
Telling a 5 year old the correct way to hold an electric saw is not optimal.
(shuddup @RoryAlsop I mean normal 5 year olds, not your uber scotson. )
If you assume that flipping bits won't go undetected then you basically assume that the MAC is useless. Then why bother ? As for ECB mode, it is intrinsically weak. Using K1==K3 is the "two-key" 3DES variant that NIST explicitly disallows. As for the overall security of your scheme, my overall professional assessment is that it totally stinks. — Thomas Pornin 1 min ago
nice "professional assessment". :-)
10:51
@AviD hahahaha - admit it, you do the same with your kids.
heh I wish
not the saw, but other tools
I do get them to at least be comfortable with e.g. screwdriver, hammer, etc...
fair enough - we don't actually have any electric saws at the moment. I do want a decent table band saw though... and they will want to use it :-/
but my kids are more brainy than physical, so even that is not natural
brainy can be aligned with physically creative
ahh sqlmap what a great time saver
I would not want to exploit blind boolean SQLi without it!
10:56
@RoryAlsop yeah but they aint
buncha lazy potatoes
pfffft
reminds me of that joke "I am in shape. It's just that the shape is... a potato"

« first day (1888 days earlier)      last day (3290 days later) »