12:19 AM
@Griffin I hope you washed your hands afterwards.

12:50 AM
I come in here early in the morning and see @Adnan advocating shameless self-plugging.
Dammit man! What's wrong with you??

@TerryChia Get ready to write down the list
Damn it! What am I doing?!

Refactoring JavaScript for 12 hours straight with Bella Ciao looping in the background

I'm not a big fan of the Reds, but at this point of the night/morning, I don't mind dying as a Partisano.

1:47 AM
@Adnan No not yet! I probably will tomorrow. Was it better?

@Simon A bit, yes.

Good.

@TerryChia is there a national holliday coming up or something?

2 hours later…
3:58 AM
So silent

4:26 AM
Kali?
Also being silent is fine if there is something to hear

@Griffin It's more about the slogan.

:P
Ik

1 hour later…
5:56 AM
Herru.

1 hour later…
7:14 AM
@avid yeah you can download a presentation for offline presenting with the free version. It has a prezi logo on it (which I think you can remove with the paid for version) but apart from that is completely free AFAIK...

4 hours later…
11:11 AM
There is magic on the streets of Ljubljana http://instagram.com/p/cY0Xp8jjp5/
Ok how do they do that?

@Raymond hehehe good one, didn't think of this :)
that "chair" part on top... would they be foldable / retractable? I mean, how do they then go off it without disclosing the trick?

11:37 AM
@TildalWave I guess that is just two of the hidden chairs cascaded together. Lots of welding and high-guage steel.

There's something so satisfying about taking a dump at work.
literally getting paid to poop.

@lynks yes, that's what I thought too... probably some umbrella like patent to fold it really fast when they need to with a click of a button / lever

12:02 PM
Gentlemen and/or ladies. I expect to chat with you guys again in a few days. If anyone needs me feel free to email or gtalk.

@LucasKauffman Moving. I'll be picking up the truck in about 1.5 hours.

@ScottPack Surely we can just communicate via Scottpackets? ;]

@ScottPack where you moving too :o?

12:08 PM
2 hours by what transport method? horse? ballistic missile?
ballistic horse?
A wild Rory appears!

Afternoon all
Good travels @scott

you at 44con this year?
I forgot if I asked you yet.

Don't think I can - I was looking at it as an option but timing seems bad

12:23 PM
:(
I had a panic moment this morning - thought I was on-site during 44con. then realised I was looking at the wrong month.

Morning gents.
@ScottPack North? Getting closer to me, I see.

1:00 PM
I just read the 'Library of Babel' short story
And as a result ordered 'the unimaginable mathematics of the library of babel'
I remember reading something when I was young that applied the same thinking to pictures. Take a standard digital bitmap, say 1920x1080x32bit color. Now generate all possible combinations.
And you have an image depicting, perfectly, your own death. And images of landscapes from distant galaxies...etc...
I need to lie down.

1:13 PM
yeah, someone actually built a bit of software that makes those images
a lot of it relies on map reduction
and even if you look at, say, 640x480 as a resolution, it's a ludicrous number.

@lynks One of the greatest shorts I've ever read.

@Polynomial that was one of the first programs I ever wrote. I was like 11, and I decided on 100x100 with 2 colours (black and white)

heck, imagine you take any one 640x480 image - you've got 235,008,000 possible versions of it where one single pixel has had its RGB value changed to something else.
and to the naked eye most of them will be visually indistinguishable from each other.
and even doing the most rudimentary visual classification at that point becomes infeasibly cumbersome.

@Polynomial yep
I quite quickly realised that my program wasn't so useful.

even 10x10 RGB is ludicrous
1.14819e+231 possible images :P

1:22 PM
@Polynomial 10x10 16colours is no better; 2.58 * 10^120

2x2 pixel RGB should be feasible though, right!?!?!?...
nope. 4,195,020,000 possible images
now the more fun part is when you do it with polygons. assuming each polygon can have one of 16 colours, and each vertex can sit on one point in a 10x10 grid, how many polygons can you map before it becomes infeasible?
for extra points, take into account that no two polygon vertices can be on the same point
for the same polygon, that is.
well it's not a polygon if two of its points are the same
it's an infinitesimally thin line.

@Polynomial a 5x5 grayscale
Would it be feasible?

probably not.

You can get a lot with a 10-step grayscale, right?

@Adnan 5x5 with just black and white (2 colours) is 33.5 million images
Reminds me of trying to convince my friend that a 6x6 grid of black or white squares is enough to give each living human a unique representation.

1:32 PM
More GPS spoofing fun. theregister.co.uk/2013/07/29/…

@lynks I can understand your friend. It sounds very unintuitive.
@Xander We really need a PKI with pre-installed certs for the civil GPS
Quite an interesting experiment, though.

@Adnan you're right, the signals really should be signed

@lynks Damn it! Replay attacks!
Screw you, one-way GPS!
But seriously, how would you prevent replay attacks in a setup like the civil GPS system?

(I have a feeling that we discussed this before)
@lynks Oh, that.

1:40 PM
@Adnan Like anything you don't know off-hand, the answer will be "magic unicorn dust".

@TerryChia I wish I had some
I remember that Thomas once talked about this, but I still don't understand how clock synchronization happens between the GPS satellite and GPS receiver.

The GPS receiver does not have a clock synchronized with the satellites.

@ThomasPornin But then how does the receiver know how long did the signal take to arrive?

It cannot have a clock with the nanosecond accuracy anyway.
The receiver does not know how long a signal took to arrive. What the receiver knows is how longer a signal took to arrive.
The receiver listens to two satellites and measures the delay between the two signals.
So the receiver is not synchronized with the satellites, but it assumes that the two satellites are synchronized with each other.
This is why a 3D position needs four satellites.

@ThomasPornin Well, it needs 3
the other point is not on earth.
right?

1:49 PM
one minute...

@Adnan we're not talking about triangulation, where you have a sphere, then a ring, then two points. We're talking about relative distance...it's a bit weird.

@ThomasPornin Ooooooh! Makes much more sense. Thanks!
@lynks Aha! I think the reason I thought it needed 3 was because I was thinking about it as if it's measuring the delay of the signal rather than the difference in the delay.

@Adnan When you have 3 satellites, you have two independent time differences (the third gives no extra information), which is enough to define a line on which the receiver is located.

@Adnan Yeah triangulation needs absolute distance

When the receiver is on a boat, it can assume that it is at sea level, and infer the exact position as the intersection of the sea level with the line.
But that's not accurate because it happens that the sea level is not exactly at sea level

1:55 PM
@ThomasPornin Yup, I think that's where the fallacy in my logic occurred.

with four satellites, you have 6 pairwise time differences, but only 3 are independent of each other (you can compute the 3 others from them). 3 independent measures are sufficient to pinpoint the location in a 3D space.
Extra satellites can be used to make the position more accurate (because of measure approximations and irksome fluctuations due to refraction of signal in high atmosphere)
Bottom-line: signatures and timestamps won't help.
Because in order to fool a receiver, you don't have to forge a fake signal, you just have to delay a genuine signal by a few microseconds.

@ThomasPornin Well that's sucks. You could have just stabbed me with that one in the beginning.
@ThomasPornin And you do that by controlling the atmosphere?

@Adnan The receiver is a machine, not a human. It cannot know if it really is on board a flying drone, or down in a lab basement and wrapped in tin foil.

@Adnan interestingly a point I noted with a wireless scanner the other day. the GPS was refusing to get a lock and after a while I realized that it was because the date was set wrong....

@ThomasPornin I'm trying to get a picture in my head. Relative distance from two points puts you on the surface of a 3D hyperbola of some kind I think.

2:01 PM
@RoryMcCune Well that shouldn't happen, no?

I saw a cool talk on fooling GPS a while back the scary bits were that financial systems are using GPS not for location but for accurate time signals (which makes spoofing dangerous for other reasons) and that apparently military systems use civilian GPS to get the initial lock 'cause their stuff takes too long to get a good lock..

@lynks The actual surface is a bit more complex because of atmospheric refraction and parasitic effects from Einstein's relativity.

@ThomasPornin But in the case shown in that experiment (a boat in the middle of the sea and an outside attacker), how would you delay the signal?

@Adnan yeah I'd forgotten that time was that important to GPS so when the phone had a time that was out of sync by a year+ it didn't work...

From both Relativities, actually, which is kinda cool.

2:03 PM
And wouldn't signing + timestamping shave off a good portion of attacks?

@Adnan In that experiment, they just need to beam a stronger signal. The satellite is far (hundreds of kilometers at least) and powered from solar panel, so it cannot be that powerful. A direct beam from a plane or another boat 10km away should have no problem overriding the genuine signals.
And one fact of the genuine signals is that they are broadcasted, so everybody receives them.

some interesting thoughts on prevention here. thenakedscientists.com/HTML/content/interviews/interview/…

It is thus "easy" to receive the signals, and beam them immediately to the victim, but with slightly altered delays between them. If done properly, the signals as received by the victim will be late by less than a millisecond, and they would have a hard time detecting that.

@ThomasPornin Ahaaa! And because GPS depends on the delays, timestamping for verification is silly here.

@Adnan Yes. With some sort of timestamping you could detect if you are being fed signals from last year, but not if you are given signals from last millisecond. Signals cover less than 300km per millisecond (speed of light and so on) so getting "late" signals (of a few milliseconds) is normal.

2:09 PM
@Adnan only because it is possible to incorporate a delay that is significant in terms of interpretation, while being insignificant in terms of signal propagation.

If a receiver has an embedded clock with nanosecond precision and accuracy, maintained synchronized with the GPS clock through some unspecified means, then the receiver could measure absolute distances with satellites and thus detect some kinds of fool play (assuming that the signals are timestamped and signed, which they are not at the moment).

@Adnan Want a bit of a mindfuck? Try to work out if Pi constitutes a Library of Babel.

@Polynomial That one is an open problem. We don't really know, although we suspect that it is true (i.e. every possible sequence of bits appears somewhere in the binary representation of Pi).

@Polynomial There was a Math.SE post about that

2:14 PM
@Polynomial This comes back to a discussion we had last week about whether Pi is normal

@ThomasPornin that is the correct answer.

161

I came across the following image, which states: $\pi$ Pi Pi is an infinite, nonrepeating (sic) decimal - meaning that every possible number combination exists somewhere in pi. Converted into ASCII text, somewhere in that infinite string if digits is the name of every person you w...

ah, interesting.

@lynks Oh yeah, that one

"A library that contains all books, contains no books"

2:15 PM
that's a thought... is there any published work that demonstrates the entropy of pi?

@Polynomial if you start using Pi to generate key material, you'll be sure to tell us won't you.

I mean, surely the probability distribution that any chosen base-n unit of pi is something that has been studied?
for at least values of 2, 8, 10, and 16

Oh no, not Pi discussion again.

@Polynomial No, Thomas hasn't talked about that yet.

@Adnan Are you implying that @ThomasPornin needs to publish his work for it to be accepted as fact???

2:17 PM
@Polynomial There is a lot of published work which basically amounts to: "We looked real hard for biases or other fast means of computing the digits, but we found none. We don't really know why."

@TerryChia No, I was implying that whatever Thomas says constitutes a published work.

@ThomasPornin so it's normally distributed in all bases >=2?

@Polynomial As far as we can measure, yes.

interesting.

2:19 PM
But infinity is a bit long; we don't have sampled a non-negligible part of it yet.

that's another irrational weird number like pi

@Polynomial wait, 'normally' has meaning here.

@lynks that's the meaning I used.

@Polynomial thought so, just checking...I expected 'uniformly'

2:20 PM
:10549185

but I see where you were going now

@Polynomial It is more than irrational; it is transcendent.
I think its status is more or less similar to that of Pi. But it has been less investigated because it is less widely known in the general public.
You can obtain a grant for computing trillions of digits of Pi, not for trillions of digits of e.

All real numbers are irrational. At least to within very accurate definitions of "all".

@Adnan Technically, everything that we write on a public medium, including the DMZ, can constitute a "publication" for purposes of researching precedents in patent litigation.

@ThomasPornin Both are transcendental

2:24 PM
@ManishEarth There are a lot of transcendental numbers, including Pi and e, and an uncountable infinity of others.

@ThomasPornin and it's being peer-reviewed.

@lynks The interesting thing is that while no bijection exists between the two sets, there is a rational number between two irrational numbers and vice versa

@ManishEarth You lost me at bijection.

@ManishEarth that I did not know...

@Adnan What ? Peers ? There are other bears out there ?

2:25 PM
@ThomasPornin I know. I thought you were using its status as transcendental to justify the grant sentence. My bad

@TerryChia means you can map each set onto the other.

@Adnan There are no equals to The Bear.

@ManishEarth So it's...fractal...always alternating between irrational and rational.

@TerryChia You can't write rationals on one line and irrationals on the other and make a one-one map between the two

Speaking of bears. This must be explained

2:26 PM
@lynks No, it isn't, because if it was then there would be a bijection :P

@TerryChia Rational numbers are countable, meaning you can define a process which goes through each number, one after the other, and such that every rational will be reached at some point.
Irrational numbers are not countable; they are a "bigger" infinity.

@ManishEarth Well 'alternating' has no real meaning when we are talking about infinitesimals :P

@TerryChia on the other hand, I can map even numbers to all +ve integers (f(n)=2n), or natural numbers to rationals
Both are countable infinities
aleph-zero

Actually Transcendental numbers are not countable, but non-transcendental numbers (aka "algebraic") are countable (they include rationals and a lot of irrationals)

@lynks exactly
@ThomasPornin yep

2:29 PM
@ThomasPornin This makes sense. Rationals are countable, so rational polynomials are countable, so non-transcendental numbers are countable.
or something like that.

@lynks That's the gist of it.

I'm definitely an irrational polynomial.

Cornerstone of the argument being that a polynomial only has a finite number of non-zero coefficients.

@lynks The question ins, is @Polynomial countable?

@ManishEarth THERE CAN ONLY BE ONE

2:30 PM
@lynks It's not quite that straightforward, because here again you have an "infinity of infinities".

@Polynomial you must have an evil twin, your reciprocal perhaps.

Countability of a root of a polynomial of degree n is mappable to countability of R^{n+1}

@lynks My evil twin is definitely not listening to Ted Stevens - Series of Tubes

But for all polynomials, that adds another infinity

THE INTERNET IS PROVIDERS
THE INTERNET IS CONSUMERS
THE INTERNET IS YOU AND ME
THE INTERNET IS NOT A BIG TRUCK

2:32 PM
@ManishEarth given two countable infinities, can't you count all possible pairings using diagonalisation. like we do with the rational numbers?

@lynks Yes, but here n can go to infinity too
R^n is countable

@ManishEarth so you diagonalise the first two. then diagonalise the result with the third, etc...?

is {x:x belongs to R^n} countable? Where n can be any number?
You get a structure where the first floor is 1d, second floor is 2d, 3rd floor is 3d, and so on
But you can map it to rationals with some trickery

@ManishEarth That sounds like an acid trip waiting to happen.

Think about encoding an arbitrary polynomials with rational coefficients (think of rationals as p/q, not recurring decimals) as a single number. It's a computer problem.

2:36 PM
arguably you're just converting bases

@ManishEarth It might be simpler to envision walking the numbers as a process.

@ThomasPornin walking? O_o
@Polynomial as in?

You can "walk" all integers (positive and negative) by doing a 0, +1, -1, +2, -2, +3, -3,...

@ThomasPornin Yep. Sure

it can be seen that for any given integer x, you will reach it at some point.

2:38 PM
You can walk all rationals similarly
or triples (p,q,r)

@ManishEarth well if you're encoding a polynomial as an integer, that integer is just re-encoded as a set of "on" or "off" gates... what's to say you can't just encode each coefficient as a different base.

or any number in R^n

depending on the size of your coefficients.

Then, if you can walk set A and set B, then you can walk all combinations with the "triangle walk"

@Polynomial I still didn't get you

2:39 PM
so you can walk pairs of integers.

so if your coefficients are all in the range 0<=n<256 then each coefficient is a set of 8 gates.

@ThomasPornin this is what I was trying to say with all the diagonalisation talk

@ThomasPornin Yes, Cantor diagonalization
Like I said, you can walk R^n

and you can encode a polynomial with 6 coefficients as 48 gates.
or a 6-byte intege.r

@Polynomial Ummm, rational coefficients

2:40 PM
it's still a question of encoding
at some point it's a set of gates that are on or off
and you can encode them as you will

Math city, math city.

Soviet Nyan Cat.

@ThomasPornin ^ walking that is not as easy as walking R^n
@Polynomial Oh, I see now. Yes, it's possible, and pretty easy to come up with a specific encoding

@ManishEarth You just have to use the degree as another dimension (more or less...)

@ThomasPornin You can't

2:42 PM
That is, you consider that all polynomials of degree n are in the big ordered list Ln.
You walk one step of list L1
then one step of L2 and one step of L1
then one step of L3, one step of L2, and one step of L1
and so on

@ThomasPornin If I use the degree as a dimension the number of total dimensions becomes infinite, because I have to allow for the remaining dimensions to be accomodated

this will make you reach any given point on any list "at some point"

Hmm
Yeah, that works

it is the triangle walk again

Much more elegant than my solution (encoding)

2:44 PM
each list is an horizontal line, and you do the diagonals
@ManishEarth You are using R instead of Q, there (minor point, but could be source of confusion for the unwary).

@ThomasPornin oops
So used to using R^n for tuples that I forgot :P
fixed
It's interesting how i.stack.imgur.com/00tp5.png is countable but not R (which can be written as Q^\infty if you cheat)

@ManishEarth Well, yes, you cheat. The point is really that a polynomial has only a finite number of non-zero coefficient (which explains why a polynomial can have a well-defined degree).
The usual proof of non-countability of R implies defining a number with an infinity of non-zero digits, and that's a crucial part of the argument.
(Not that having an infinity of digits is really a problem for countability, but having an infinity of independent choices for digits does it.)

Is anyone good at charsets and all that junk? In the header of a page, data from a database is fetched and displayed properly with accents. However in the body, data is fetched but invisible if there is an accent in the word.
It is not the same table but the collation is exactly the same. I'm confused as hell.

@Simon With HTTP, things are complex
because the charset can be specified in the HTML itself, and in the HTTP header, and the browser can enforce its own default value
If you don't use UTF-8 everywhere you only get what you deserve.

@ThomasPornin Yeah. There's a software relying on the very same database and there's no accent issues.

2:53 PM
@Simon The golden rule is: Charset in the dataase should be UTF-8, fetch using UTF-8, send the HTTP headers as UTF-8, set meta to UTF-8, and set your IED to UTF-8

@ThomasPornin Thing is I haven't built this application, I'm migrating it to a recent version of MySQL. And of course, they haven't used UTF-8.

@Simon Then export and import again as UTF-8

@Simon Look at the bytes on the wire to see what appears in the HTML page, and also the HTTP header.

@Simon I'd guess that the BOM is what's breaking it.

@Polynomial Oh! PHP and BOM!

2:55 PM
(inb4 "there's no BOM in utf-8"... yes there is!)

@Adnan I'm gonna have to manually change web pages, too. They have a class for headers but they didn't put the HTML tag in there. No idea why.

@Polynomial If there is a UTF-8 BOM, the browser will infer UTF-8 and be very sorry when it encounters non-UTF-8.

@Adnan don't get me started :P
@ThomasPornin in some cases (see Adnan's point) it breaks things regardless.

@Polynomial Well, yes, though it appeared late in the show. Microsoft lobbied a lot for its inclusion.

Webpage: charset=ISO-8859-1 DB: latin1_swedish_ci
Great, just great.

2:57 PM
@Polynomial Well, yes: if HTTP header and HTML contents (BOM, meta tags) don't agree, hilarity can ensue. But with dynamically generated content, the Web server cannot necessarily adjust the header to match the HTML (Apache will do the match for static files, though).

*nod*

@Simon my condolences

I still don't get why in the header the accent is displayed but not in the body.
Also, if I add a static value with an accent in the body, it is displayed. It's a DB issue but still, I don't get it because the first value came from the DB too.

you know what scripting languages need? a way to trace which functions and lines of code have "touched" a variable between two breaks.
that'd be so frickin' useful.
and should technically be quite simple in managed languages like .NET, where all operators are technically function calls.

Oh man. Fuck.
Does MyISAM have anything to do with charsets?

3:02 PM
sure. collation.

Because I switched to InnoDB.
It used to run on MyISAM.

every time I watch this, I want a t-shirt with his face on the "YOU CAN'T TRUST THE SYSTEM!" bit, to wear at conferences.

I think I'm just gonna have to go through all the damn charsets and switch to UTF-8.

@Polynomial this is an excellent idea.

@Simon InnoDB has collation too.

3:04 PM
@Polynomial Or is he?

@Polynomial MAAAAAN
@Polynomial Are you PART OF THE SYSTEM?!

@Polynomial some poser hands me cake at a birthday party

I used Tor, but I didn't like it what I found there... you know what I'm talking about...so there's no Secure VPN? it doesn't matter if you pay or not? — jcho360 14 mins ago
JESUS!!!

3:09 PM
I hate those people who go "Oh I don't use Tor because it has child porn"
Well don't fucking use the Internet then!!
Have you seen the Internet and what it has?

I'm definitely vectoring that screenshot tonight and making a t-shirt.

@Adnan I assume he means the hidden services protocol.

Rule 34. 'Nuff said.

@DavidFreitag Rule 34 has ruined so many aspects of my childhood.

@lynks Well if you're trying to find CP you 'll find it anywhere. Tor or not.

3:10 PM
the pinnacle of which was a picture of Worf having sex with a flamingo.

@Polynomial Why, you found Twilight Sparkle on there or something?

@Polynomial Rule 34 has made me turn away from the internet for weeks at a time.

@Iszi I find that just through going to Facebook.

@Polynomial Heh. That reminds me of something. One sec...

I'm not phased by furry stuff. That stuff is mild compared with some of the weird shit out there.

3:11 PM
The ED used to have a page, I think it was called the most offensive page on the internet.

one of the worst ones I've seen... picture of Misty from Pokemon, cut in half, her bottom half being raped by Blastoise, her top half impaled on a wooden spike.
at that point you really question who got up in the morning and thought "yup, that's what imma draw today."

Crusty Cunt Flakes? I can never eat corn flakes again. Thanks 4chan.

Well, crap. Kinda hard to find this one safely from work. I was thinking of an LICD strip.

most of LICD is NSFW
I just about manage to get away with QC and GWS at work.

@ThomasPornin I know, I know. But on it's face it's counterintuitive.
Once you've got a bit more maths intuition it becomes obvious.

3:18 PM

Personally I gave up on the "infinity" related corners of maths
I'll stay in simple happy "arbitrarily large but finite" land :)

@CodesInChaos But why, that's when math finally gets fun

@CodesInChaos That land is called "physics".

@ThomasPornin at least, assuming a closed universe...

@ThomasPornin Well physics is weird in its own way. A typical case: "We assume x, which is wrong, but that doesn't matter. It'll work out anyways."
somewhat similar to the random oracle issue in crypto
another fun part of physics is notation. Each corner is using completely different notation, often in ways that ca only be parsed with sufficient contextual knowledge.

3:26 PM
@CodesInChaos definitely agred on that one.

@ThomasPornin and by closed I might mean open.

@CodesInChaos It's actually "We assume x because not assuming it is too much work. Well worry about the perturbations later.". But yeah :P
@CodesInChaos That is true. Even particle physics reactions use a different notation between nuclear physicists and particle physicists and chemists

for extra fun physicists overload the meaning of a single symbol within a single text or even formula

@CodesInChaos ooh ooh you mean eigenkets and eigenvalues (eigenbras too I think). Meh, that's not so confusing

3:42 PM
@CodesInChaos Wait! You're not a mod?
You're always so moddy
In a good way

@Adnan You said you researched VPN providers for privacy and didn't find a decent one, what about VPS?

@Simon In my "research" I've only looked into service providers who explicitly offer VPN service.

@Adnan Oh. You could set up a VPN on the VPS but I'm assuming that it would be the very same situation and probably more expensive too.

@simon are you talking about the privacy vpn question?

@GreenFly Yes.

3:51 PM
I have used dreamhost and futurehost and have set vpn on my servers,both of the companies keep logs and provide LE with your information.

@GreenFly Are they both US-based?

yes they are

@GreenFly Well no wonder then. All U.S. host services have pretty much the same policy.

Wish the Cayman Islands had VPN services..

@GreenFly That sucks.

3:57 PM
If they did what they do for money with my internet traffic, that would just be... fantastic.