« first day (1678 days earlier)      last day (3500 days later) » 

 
5 hours later…
05:27
TIL my laptop had an nvidia quadro 1000M which I hadn't installed the right drivers for (on linux)
tfw
so if this laptop was fast, it's going to be awesome when I give it the right drivers :D
Adi
Adi
05:41
@Ohnana I hope that you see that the whole nomination is a joke.
I have, repeatedly, said that I'm not interested in being a mod.
I will cause a lot more issues than the ones I'll be fixing.
I'll close questions just because they don't feel right to me. I'll suspend users for saying something that can be seen as not nice, all while not being nice myself
@Adi Nobody is perfect.
Adi
Adi
@DavidFreitag This is not about perfection or anything close to that.
I do NOT have any of the qualities to be a good mod
This is not self-hate
This is how it actually is
lol
@Adi: IMO, the most important qualification for being a mod is being sick and tired of people bugging you to stand.
@Adi Don't worry, it was a joke. I too lack the proper qualities to be a mod.
Adi
Adi
See how @RoryA usually waits to be the 5th closing vote so he gives the community a chance to decide? Yeah. I won't do that
05:46
@Adi meh. Community can reopen ;p
Adi
Adi
Anyway. I still have the same opinion I've expressed several times here. I don't have an interest in being a mod, and I think I will have a negative effect on the community if I become one.
@Adi funny thing is someone won a mod election on a similar platform
He did it.
06:03
And was thinking, that wouldn't that serve are a good way to identify a hacker?
The question talks about implementing a delay between sending the response for an incorrect password request
So if we receive multiple parallel bad requests, we can be sure that it's someone not playing by the rules, not waiting for the pre-defined interval between login attempts, and thus mark them as malicious
And the alternative for them would be to wait between login attempts, which would slow down their brute-forcing.
Thoughts?
06:38
@PeeyushKushwaha This is a pretty common technique for identifying attackers. Check out products like fail2ban for examples. It's not perfect, and the false-positive rate is abnormally high. At work I frequently mistype my password 5 times in rapid succession just because it's so damn difficult to type correctly with all the case changes special characters.
But if you set your threshold high enough (say 15 attempts in 2 minutes) then you'll get a better result.
the problem with multiple parallel requests is that some accounts are more commonly attacked than others. You can expect "admin" to get password attempts coming frequently by dozens of independent attackers.
So if the legitimate user mistypes his password once, the chance is high that he'll be flagged as an attacker.
because REAL attackers will have typed in bad passwords at the same time
@tylerl, even if you as a proper user mistype your password a lot, but if you wait between login attempts as the page takes time to load for you, then that shouldn't mark you as an attacker. What should mark you is if you're sending multiple requests, which can be a marker that you're not using the application UI which will definitely wait for response of the previous login attempt, whereas an attacker won't wait...
Adi
Adi
@tylerl I don't think that's what @PeeyushKushwaha is talking about
@PeeyushKushwaha Oh, good. I was gonna type that
I think I agree with @Peey
That does seem like a good way to flag a malicious user
If you get multiple login requests from a single IP in a short time, that's a good sign of an attacker.
But it's also a good sign of a public VPN endpoint
or a NAT gateway
Adi
Adi
06:44
@tylerl :|
Nope
Oh right, a VPN endpoint would be a "false positive" with this method
Adi
Adi
I doubt that they will be attempting the same username
Why would it be?
@Adi You have 10 attackers an 1 legitimate user all attempting the same username at the same time
all but one is an attacker. Which one's the good guy?
Adi
Adi
@tylerl So you're narrowing it down even further
To a quite unlikely scenario
the attackers and the legitimate user being on the same VPN/NAT
Yes, it's possible
No, it's not likely
It's not a BAD signal, but it does treat specific users harshly, and often these users are high-profile, often celebrities of a sort
Adi
Adi
06:46
It's far from likely
@tylerl That's why you increase his "badness" score
and not ban him
Hmm. Even if you restrict it to one user per IP, then one attacker could potentially try to hit multiple users instead of one. E.g. try to login from different accounts using the password "password"
@Adi it's far from likely on your blog. Less far from likely on twitter
But you could use other markers in addition to this to prevent that
@PeeyushKushwaha yup. and that's pretty much exactly what companies like twitter do.
Adi
Adi
Which is what you do with that information. Like I said, increase his "badness" score
If you notice this behaviour over a long period of time, the badness score becomes too high
06:48
typically captcha.
Adi
Adi
If it goes over a threshold, you introduce extra throttling
Yeah, captcha for example
In the end, you'd have a banning threshold
Hmmm.. aren't there 3rd party services that do that for IP addresses?
Similar to spamhaus
Yeah so it seems like an effective system to me. This in addition to captcha if an ip attempts to access more than n number of accounts, and that too, with significant failure rate
Typically with Google, if your IP gets marked as "bad", then users will get a captcha. If they solve the captcha, they get a cookie which means they won't have to solve the captcha again.
Adi
Adi
@tylerl Indeed. I've noticed that
However, I've also noticed that with really bad IP addresses even having solved a captcha wouldn't help you
Is that true?
(Google case)
@Adi each product gets to make their own decision on how to proceed. I've seen cases though where users were simply told that they were blocked.
Adi
Adi
06:53
@tylerl Yup .
I've had that with Google Search
when being on really bad exit nodes
because you were running a bot, amirite?
ah. tor. yeha.
Adi
Adi
Hmmm.. how is captcha pronounced?
I've heard from PLENTY of companies who have complained that nobody at their office can use [insert product] because they get some "you are running a bot" message. The only real advice I can give is "well, find the bot then."
"CAP-CHA"
@adi I've always thought it's pronouced cap-tcha
Adi
Adi
06:55
So, kaptsha?
@Adi yarp
Adi
Adi
I've heard it once Kaptshka
Adi
Adi
Maybe it's a cutesy name or something
should there be sh in there?
Adi
Adi
Alright. So, less T
Interesting, if you hit the listen button twice, it utters slightly different pronunciations
@Adi Well, it should be said that the "American accent" is notorious for not pronouncing t's.
"PTCH" is too many consonants to pronounce all of them.
so the T gets left out
cause its in the middle
Adi
Adi
Anybody used Security Compass SD Elements?
06:59
@Adi nope
also, if anyone wants to work at Google, our team needs people. And it's tough to staff.
@tylerl interesting
Oh - this is happening:
@Xander who needs rep when you've got beer?
@RoryAlsop our team has an Dublin office. It's not scottland, but it's the same timezone.
practically walking distance. And swimming.
lordy. I better go to bed. night!
@tylerl night
07:27
@tylerl amusingly when my missus' ex-company moved their European base to Dublin they thought it was in the UK (d'oh) and were surprised when UK staff were reluctant to move country to be there...
morning all
Morning.
@RоryMcCune That is actually a really quick and easy trip from here
The worst bit is the taxi in from Dublin airport :-)
@RoryAlsop indeed but moving there is a big change as it's a diff. gov..
@RoryAlsop yeah I've done that run, not fun or cheap
@RoryAlsop heh I now have a badge on workplace to go with my points for answering one question
@RоryMcCune got enough to vote there? I've not been very active over there yet
@RoryAlsop I've got 241 so not enough to nominate if I had felt inclined.. can I vote with that amount?
@RoryAlsop and I'm surprised you've not done more on that site, you'd think it'd fit in well with your HR work :op
hey there
07:38
'morning
(or whatever)
Adi
Adi
Morning!
Bonjour!
Huomenta!
صباح الخير!
בוקר טוב!
It /might/ be the afternoon for some, you know ?
Or evening
@RоryMcCune you can vote, yes
@Stephane crazy talk
:-)
we all know Zulu time is the only true time
@RоryMcCune Well - I got busy
@RoryAlsop It's nearly 4am here
1 min ago, by Rory Alsop
@Stephane crazy talk
:-)
07:47
@DavidFreitag See ? Night owl happens
It's nearly midnight in California.
@Adi I finally got around to reading your nom :-)
very good
@RoryAlsop On freenode, they got this thing where they consider it's morning when the user connects or become active, and it's evening when he leaves :)
feels to me like this one isn't actually a dupe. Thoughts?:
26
Q: Dangers of downloading without opening file

ahinathIf we download a file from the Internet, but not at all open that file (if we keep autorun off), are there any chances of virus spreading from that file?

@M'vy Or should we use URT (Universal Rory Time) which suggests that it's whatever time Rories say it is? ;-P
Or UBT!
No-one would dare challenge UBT
Universal Bear Time
@RoryAlsop A wino would.
07:53
@RoryAlsop Well I agree the answer is not 100% dupe and it could be good to have a more generic one, i.e. any tools can get exploited..
The other one is looking at the well known risks from going to the website - a drive-by attack, whereas this one assumes the file is on disk. I think that's a separate scenario.
@RoryAlsop actually the CW answer could be improved to become more "canonical"
That's weird I did not see that answer at the time.
They should really show the answers when reviewing...
(I mean not just the answer count)
hm that's weird. My raspberry-pi do not accept my connections from outside my home-net
Ahah found it. The home router did not reset the port forwarding for some weird reason :)
08:46
Xander and Schroeder leading
I got 22 downvotes and 56 upvotes
Ooh I finally got that :)
> You've earned the "Strunk & White" badge (Edit 80 posts).
never too late to improve :)
@M'vy yay
@RoryAlsop never too late to improve candidate score :)
ahah
Damn I forgot how long all these candidate questions takes to answer properly :P
09:11
@M'vy lol - gotta work for your mod swag
@RoryAlsop true
@Adi You do realize that you don't really have the highest candidate score, right? Just sayin' :)
@Stephane it's always morning in the DMZ
+/- 12 hours
24
A: The Memes of Information Security

Scott PackMeme: Canonical Time Zone Originator: Unknown Cultural Height: Early 2011 Background: American hubris on the part of our East cost users resulted in blanket declarations that EDT/EST (UTC-4/5) would be considered the de facto timezone of The DMZ. This was brought upon by inconsistent time of d...

09:26
@Xander flair play tip: you should add your name on top of your answers in the meta post :)
I have a feeling that @polynomial may not be too engaged in these elections, seeing as how he's in Japan on his Honeymoon
@RоryMcCune well I have read his nomination
09:57
@RоryMcCune wow - is that this week. I meant to send him congrats
@RoryAlsop yeah the wedding was the wknd, IIRC I saw it go by on twitter, then there's pictures of Tokyo this morning...
> SQRL master keys are always encrypted under the user's current password.
^ what's the point of it then ?
The more I read, the less I find it interesting
@M'vy lol
10:16
> A MASTER KEY SHOULD NEVER BE EXPORTED UNLESS IT IS PROTECTED BY A STRONG PASSWORD
No sh** Sherlock
> The computation to export an identity master key is designed to deliberately require 60 full seconds of intense processing of “memory-hard” functions by the device performing the export.
ANy other views on:
2 hours ago, by Rory Alsop
26
Q: Dangers of downloading without opening file

ahinathIf we download a file from the Internet, but not at all open that file (if we keep autorun off), are there any chances of virus spreading from that file?

bold isfrom me
@RoryAlsop well you have mine :)
@M'vy yep - I saw that. I think one more and I'll reopen. I do think it isn't really a dupe, but generally like the community view, rather than using my mod-powazz
@M'vy yes - exactly
@RoryAlsop well, the "idea" is a dupe. The use case is not.
@M'vy oops - I responded in the wrong line
10:19
Com'on you @RoryA !! You respond to future sentence now :PP
Don't get all Doctor W now :P
anyway the "feature" is interesting :)
can someone please reject this?
it's really annoying, post needs more than just removal of TiA
yep
thanks @M'vy
10:24
no problem
> The hope is that someone who is merely playing around (for example an innocent child) will cease guessing not wishing to cause the phone's owner undue trouble.
^ Ahah keep making me laugh :)
so all of the candidate elects would be new to moderation?
@TildalWave define "new to moderation"
as in none were mods on SE before
or pro tem
or CMs :)
Well, for my part I've already posted on the Q&A on meta
never been mod on SE before
neither have any others as far as I know
10:29
And would I've been, I probably wouldn't have nominated
or maybe even sooner then since you'd already know the ropes
@TildalWave sooner?
I mean with less hesitation, if there even was any to start with
oh.
I've nominated for multiple elections already, so with me getting more rep/badges/review the hesitation goes down :)
it's a graduated site, mod work is (or should be) pretty much business as usual here, lots of things already have established procedure and you just follow it... not the case on betas tho
if I wasn't a mod on a beta I might even nominate here
but that takes too much of my time so I couldn't guarantee to be a good mod here
10:34
@TildalWave well I'dsay the challenge are different :)
@TildalWave That was also my view on the topic
every mod I talked with about this says it's a lot easier to be a mod on a graduated site ... I'll take their word for it because I know that it isn't easy being a pro tem mod on beta.
Anyway, time to get lunch
@TildalWave betas are in constant evolution, and the community is defining itself
Definition is just one part of it, there's also community building and teaching newcomers the ropes. You get a lot more support on a graduated site from an already established community. On betas, you're looking at more questions each time you answer one. It's more like having a kid vs having a pet kind of a thing.
Boop
Adi
Adi
10:49
@TildalWave I do have the highest candidate score.
You said I don't. This means that you know someone who has higher. Who is it?
@AviD AviiiiiiD
@RoryAlsop This is quite true. Very philosophical of you to notice.
@M'vy Ah, good tip. Done, thanks.
@Adi OK what is a candidate score and where do I see it? I thought you're referring to site reputation, in which case Poly and Lucas have more of it.
@TildalWave definitely! This site is by far my easiest!
@TildalWave Go to the nominations page, and there's a blue box at the bottom of each person's nomination with their candidate score.
10:58
@M'vy exactly this - a lot of the evolution is fun work, but still work.
Once a site is as established as Sec.SE, the community does almost all the work, and mods only need to cover exceptions and a bit of housekeeping
A small beta doesn't have that core of high rep users who can carry out these tasks
so mods do everything
raz
raz
11:10
It's a tight race between @Xander and @schroeder
Nah - this race is between who is last and 2nd last :-)
@RoryAlsop @raz And that race is not so tight. :-)
raz
raz
@RoryAlsop Oh are we supposed to be voting people off the island?
@raz The primaries are to determine the top 10, who go on to the finals. Since we only have 11 candidates in total, essentially yes.
It's funny the troll submissions still are doing better than Simon
11:17
Hahahahaha
@LucasKauffman They have higher candidate scores.
They aren't real troll submissions - we know who they are.
If they were new accounts it would likely be a different story
raz
raz
@Xander Do we then narrow it down to 2 people where we have a CTF competition?
The troll accounts would be fun, and CTF would be the best way to determine a winner
@Adi oh shit how did I miss this? :))
11:22
Also, I've never written a random int -> string function before that I could safely say is sound; who thinks this is terribad? codereview.stackexchange.com/questions/97507/…
@Xander you welcome. Actually it was suggested by @GraceNote in the Q&A introduction
I wonder where I would have ended up if I had done it seriously
raz
raz
@Amelia It's pretty terrible to use Math.* for security purposes.
@raz I'm not using Math.*
raz
raz
Oh I thought you were commenting on the OP's code
11:32
I'm using RandomSource.getRandomValues in that function; the OP is the one doing that
@raz you need to read Amelia's answer - to me it looks ok, but I'm not a js person
Just wondering if int % chars.length is the right way to pick a char out of an array at random
Never actually looked at how to go from int -> string securely...
raz
raz
@Amelia Well you have an array of uin32's, if logical AND with 0xFF is an option, I'd use that.
And then the % char length seems fine to me.
@M'vy Ah, so it was. I missed that completely.
raz
raz
@Amelia I'm not a JS person does crypto.getRandomValues(new Uint32Array(length)), return a byte array?
11:38
@raz it fills a typed array (Uint32Array) with random values of that type, by reference, and also returns it (because logic)
it's an array of random uint32s of length length
You can technically use any typed array, but I just used uint32 because it makes using it modulo charLength easier since you don't have to account for negative values
Also, size. :3
raz
raz
@Amelia Yeah that seems to do what you want
In the OPs code they have randomIndex = Math.floor(Math.random() * password.length); I'm not sure how that stays within the bounds of password's length for an index... but ok.
It's also early in the morning, and I haven't had coffee yet
var buffer = new Uint32Array(), retval = crypto.getRandomValues(buffer)
buffer and retval are the same object.
Also, brb food

« first day (1678 days earlier)      last day (3500 days later) »