The h Bar

0celo7

13:00

@ChrisWhite I need TeX help!

@ChrisWhite ack, it says this is chapter zero

that's not what I want

Dang, proof is wrong.

Physics Meta

13:17

0

Q: What is a good mentality about casting votes responsibly?

I find myself casting quite a lot of up-votes on questions, usually reaching the daily limit and was wondering if I might be polluting the site with that. What are other people's mindsets about casting votes? Do you have a certain criterion other than that the question has to fulfill the rules o...

0celo7

@Slereah you around

Slereah

mb

0celo7

drive.google.com/file/d/0B4ZCX9GkmqtmcklaaDViY3RwTWs/…

@Slereah I need help fixing Lemma 7

I know it's true

I'm trying to "average" the velocity vector along the spacelike curve to get a single vector

And the exponential map of this should be the endpoint

but I'm having a hard time proving the average should be spacelike (I know it has to be)

Ah @ACuriousMind would you happen to know how to remove the 0. in front of the theorem numbers?

ACuriousMind

13:34

Remove the [section] that is presumably behind your \newtheorem line.

0celo7

That wasn't it, but it was a good hint, danke

Slereah

Is the lemma basically "if p and q are spacelike related, there's a spacelike geodesic between them"

0celo7

Special relativity is too hard

@Slereah Iff.

The iff is the hard part.

Slereah

Well one is trivial

0celo7

Correct

Slereah

13:39

So only one part is hard

0celo7

That's why I said "the converse is obvious"

Slereah

How do we know this is true btw?

0celo7

And I'm pretty sure you need iff for Theorem 8's proof to work

@Slereah Draw it

Slereah

Did you check the proof that all null-related points are joined by null geodesics

Maybe this will help

0celo7

Where is that proof?

That would be helpful, yeah

Slereah

13:40

It is in Penrose, IIRC

0celo7

Page?

I have Penrose with me

Slereah

At work m80

0celo7

me too :P

Slereah

It's in chapter 2

Dunno the page

0celo7

2.19?

Slereah

13:42

I dunno man

Don't have it on me

0celo7

let me get my legal copy...

Slereah

It is by the way what convinced me that you can't get around chronology protection easily

Seems like it yeah

Causal non-timelike null related points have a geodesic between 'em

0celo7

Ok, will have to read later

@Slereah but isn't it true

the thing I'm trying to prove

if you have an everywhere spacelike curve

then the end point has to be spatially related to the start

in Minkowsi

Slereah

Oh in Minkowski certainly

0celo7

so what I need to show is that $-(q^0)^2+\sum (q^i)^2>0$

but this seems nontrivial.

Slereah

13:48

In Minkowski can't you just prove that from a point, a null geodesic reaches every other point

I mean those are just straight lines

0celo7

the fuck do null geodesics have to do with anything

Slereah

Errr

Spacelike

0celo7

Yes, that is trivial

But what I have is $p\asymp q$

And from this I need to show that there's a spacelike geodesic connecting them

This is hard.

(yes, I made up that notation)

Slereah

Are we proving it only for Minkowski or in general

0celo7

Just Minkowski for now

Slereah

13:49

Because doing it in coordinates probably works fine

0celo7

No, it's hard!

Did you not read my thing

you get an integral over a unit square

at first I thought the integral had to be positive

but then I realized we're taking derivatives at different points of the curve

Maybe there's some continuity argument

Maybe we can bound the integral from below.

Maybe I need a bump function, lol

Somehow get rid of the parts where the inner product is negative

...that's not even that bad of an idea

Slereah

Aren't all geodesics in Minkowski space just a straight line

0celo7

Yes, but who gives a shit

Stop proving the converse, that's trivial

Slereah

Can't you check that a straight line coming from one point will intersect another

0celo7

What?

Slereah

13:59

Iunno m80

0celo7

this seems like a really easy special relativity problem

why is it so hard

Fuck it, let's do what physicists to best. TAYLOR EXPAND

John Duffield

14:21

@JohnRennie : I know why. It's really simple.

0celo7

@JohnDuffield enlighten us

David Z

Not here

0celo7

lol

John Duffield

14:37

@DavidZ : please explain why not here to the other posters.

0celo7

@JohnDuffield because he's tired of our back and forth

David Z

@JohnDuffield In this case, because you've already been told not to have physics discussions here.

Danu

-1

Q: Is there a proven formula (or pathway) which can curb chronic poverty in inner cities?

When you have rats in the neighborhood you want them eliminated. You don't go to your Legislator again and again for more money to throw at the problem again and again because we still have rats. Therefore, this question is not to encourage discussion, but to identify the proven elements in a for...

soft-question education history

@JohnRennie The entire concept doesn't make sense in Newtonian gravity

Slereah

What is the lagrangian of the inner city

0celo7

@Danu Huh? Having a body that has a certain escape velocity makes sense

@Slereah you joke, but I once tried to do this

Danu

14:47

@0celo7 But infinite velocities are possible so there are no "black holes" in the GR sense.

Slereah

Well there is econophysics

0celo7

It had a lot of terms

Slereah

But it's not really Lagrangian related

it's more stochastic equations

0celo7

@Danu Ah, ok, but if we say that there's something that light can't escape, I think that's valid

Unless it's aided, ofc

You should "slingshot" it since velocities add.

@Slereah I tried to do the path integral

Danu

Ehh yeah... $c$ is not independent of frame in this context

Slereah

14:49

This one?

0celo7

@Danu That's what I just said

Danu

I think it's definitely nonsensical to talk about black holes

Slereah

@Danu : But the frame of the source mass is the important part

Danu

@0celo7 Note "ehh yeah"

Slereah

Well they were not called black holes back then

They were called

Astre occulté

Danu

14:50

haha

Slereah

I forget the english translation

Danu

Hidden star?

Slereah

kinda

Danu

@Slereah Very encouraging prospects ;)

0celo7

@Danu oh, I didn't read that as agreement

Slereah

14:52

Yeah I think he had some pretty poor idea of the phase of good prospects versus bad ones

0celo7

What if I use the mean value theorem

Slereah

Remember the rule @0celo7

Be nice

Pick the nice value theorem

Danu

lol

@Slereah Especially the "post doc post doc post doc" lol

0celo7

$$\int_0^1 c'^\mu(s)\,\mathrm{d}s=c'^\mu(k)$$ for some $k\in[0,1]$.

HA

I think that's it

@Slereah good job

@Slereah nooo it's wrong

@Slereah The problem is that the $k$ need not be the same for all $\mu$.

0celo7

15:24

@Slereah What if I do Morse theory on the space of spacelike paths

Whatever that means.

Slereah

I don't know Morse theory

0celo7

Maybe just a variational method then

I have an idea.

Slereah

Aren't there no smallest path for spacelike curves?

0celo7

There is a smallest!

The geodesic!

Slereah

Well geodesics and shortest path aren't the same

Can't you make the path arbitrarily small by getting closer and closer to a null curve?

0celo7

15:34

Can you do that with a spacelike curve

Slereah

I think so?

I dunno

I vaguely recall that point

0celo7

Back to the drawing board then

Slereah

5

A: How does one measure space-like geodesics? Or: What is the physical interpretation of space-like geodesics?

If you have two points $p,q$ spacelike separated in a spacetime $M$ there is not anything like the shortest spacelike curve joining them! Any spacelike curve joining them can be continuously deformed closer and closer to a lightlike curve joining the same points. So the $inf$ of the set of the le...

0celo7

Bah

Danu

@Slereah You can approximate every timelike path arbitrarily well by null curves in any case

Slereah

15:45

Yeah I remember that

0celo7

By null curves?

Slereah

Zigzagging

Danu

(just zig-zag)

0celo7

Never heard that one

Danu

@0celo7 lolwut

Null curves are already null curves

0celo7

15:46

Nevermind

I misunderstood you

And you misunderstood me

And the phone doubled my message and only let me edit one.

John Rennie

@Danu it isn't that I expect the exercise to have any physical significance, because of course it doesn't. But in the way that people will waste their time doing crossword puzzles or Sudoku it might be interesting to waste my time looking at this.

I appreciate it's a form of numerology and devoid of meaningful content, but I'm just curious why the two approaches give the same result.

Danu

OK

John Rennie

BTW did you read the Ohanian book?

Danu

No, sorry. I'm quite busy with all kinds of other stuff

Slereah

Isn't Newtonian gravity basically just Schwarzschild metric with $g_{rr}$ set to 1?

Danu

15:49

My girl is moving---lots of work to do---and other than that I'm reading a book on Riemann Surfaces at the moment

0celo7

So Danu is capable of being nice, presumably.

John Rennie

Or $g_{tt}$ set to one depending on your choice of signature. I'll have a think about that when time permits. Right now I'm writing corrosion modelling software for the nuclear industry (EDF).

0celo7

Phone, please fix autocorrect.

John Rennie

If you live near a nuclear power station now is a good time to consider moving :-)

Slereah

Are you in France or just subcontracted

John Rennie

15:52

I'm a subbie. I contract for National Nuclear Labs in the UK, who run the project for EDF.

It is surprisingly good fun :-)

0celo7

@JohnRennie About 45 minutes from one.

But it's a DOE research reactor, doubtful you're working on it.

(My university maintains it.)

Danu

@JohnRennie Somehow I thought you're retired... :P

0celo7

^

Danu

With y'know... So many answers on this site and all

0celo7

Maybe he's just a quicker typist than you.

John Rennie

15:55

@Danu I quit regular 9-5 work about 8 years ago, but I started doing small coding jobs for friends and word got round.

0celo7

how do you "quit regular work"

How old are you?

John Rennie

@0celo7 55. I can answer but it's kind of real world boring stuff. However I'll answer if you want ...

0celo7

Maybe you haven't heard, but real world is what's important to me

Danu

@JohnRennie Oh, that young still?

The words "Semi retired old time computer nerd" made me picture someone ~70+

Slereah

70 is more retired than semiretired

John Rennie

16:00

@0celo7 start saving regularly and put the money into a unit trust or some other stockmarket related fund that it ins't easy to get the money out of. Do this for 30 years. Retire. Simple as that.

Slereah

Can't I just buy candy instead

and eat it

and somehow retire

John Rennie

@Slereah Yes, that'll work :-)

Slereah

Currently i wouldn't mind investing in a house

Rent is bloody expensive and not building much equity

John Rennie

If you're in the UK I'd be inclined to wait and what effect the tax changes on buy to let properties will have. I'd say there's a fair chance they will depress prices a bit.

0celo7

Assuming I get my parent's house one day I'll have a very nice investment. The area is exploding, housing will go through the roof.

But I plan to marry rich anyway

That's a good plan, right

Slereah

16:06

Not in the UK tho

And I'm not buying it now anyway

John Rennie

@Danu old time computer nerd - I started programming on an IBM 370 mainframe in OS360 assembler. It's astonishing, and I mean that literally, how much being a computer nerd has changed in forty years.

Slereah

Won't be getting a loan with my current job

@JohnRennie Yeah I kinda wish it was still like that

I really hate programming high level

John Rennie

I'm not sure I miss programming in assembler. My favourite language remains C++, which isn't that much different to assembler.

Slereah

Yeah but I'm doing Javascript :(

David Z

@JohnRennie probably the first time I've ever heard that said

John Rennie

16:09

I hate Javascript.

David Z

I mean, I get it, but it's a very rare sentiment

Slereah

Then you are a wise old man

0celo7

That's why I don't have JS

John Rennie

@DavidZ C++ is such a clean language.

Slereah

I prefer C#

C# is much cleaner

David Z

16:10

@JohnRennie That is the opposite of my impression of it

granted, "clean" in the context of programming languages can mean very different things to different people

John Rennie

When I write C++ I have a pretty good idea of the compiled machine code is going to look like.

When I write C# I have absolutely no idea of what's going on under the hood.

David Z

@JohnRennie Ah, yeah, that I get

Slereah

Oh yeah

The C# compiler is fuck huge

John Rennie

C# is the future though

I have done C# work and it is easy to code in because the standard libraries are so good.

And you don't have to be careful with pointers :-)

0celo7

What are you people talking about?

Slereah

16:13

$C\#$

0celo7

What's that

Slereah

A programming language

0celo7

Is that some weird differentiability criterion

John Rennie

@0celo7 :: John faints away in horror ::

Nerds just aren't what they used to be

0celo7

Not a nerd.

Slereah

16:14

When my C# professor described it a bit I said to him like

0celo7

Only nerds know about computers

Slereah

"Oh you mean those things that Lisp could do in the 60's"

John Rennie

@0celo7 Guilty as charged

0celo7

I'm not a nerd, so I don't know about them

John Rennie

@Slereah I remember Lisp. I write a program to calculate pi to 500 decimal places in Lisp just to prove it could be done.

David Z

16:15

@JohnRennie I would imagine the stereotypical nerd as a language purist in some sense - they'd probably code in C and Lisp and honestly not have any idea about newer languages like C# or even Java

Slereah

@JohnRennie Well you can obviously

I don't know about execution speed tho

David Z

That being said I'm not overly happy with the whole idea of this being a chat room for nerds anyway, but whatever

John Rennie

@Slereah it was slow :-)

Slereah

I'll bet

0celo7

Slow?

Like a few seconds?

Slereah

16:17

machine steams for 20 minutes

$3.$ displays on screen

0celo7

Wtf

My phone could probably do that instantly

Slereah

your phone isn't an old timey tincans computer

John Rennie

I can't remember, but in those days you ran anything big as an off-line task. So you started the job, went off for a coffee then came back to what creative way it had crashed in this time.

Slereah

heheh

I had programs like that when I worked in big data machine learning things

0celo7

Sounds like when I try to write stuff in TeX

John Rennie

16:19

I think an IBM System 370 was about 1Mip

0celo7

What

@JohnRennie I need SR help

John Rennie

@0celo7 ?

ACuriousMind

...will you ever be able to do anything without soliciting help from everyone appearing in this chatroom

0celo7

No.

Danu

(spoiler: Nope)

0celo7

16:22

Oh fuck it, damn phone

John Rennie

@0celo7 you will buy an iPhone

Danu

No, I will not "fuck off", @0celo7 :P

0celo7

@JohnRennie I have an iPhone

It's no longer autocorrecting

@Danu sadly I think that's the case

Slereah

I have a zPhone

John Rennie

@0celo7 my point exacty. You should get a phone that works.

Slereah

16:24

Proudly made in Moldova

John Rennie

Anyway, what did you want to know about SR

I can explain anything except for geodesics in Rindler coordinates

0celo7

@JohnRennie no, I don't want to give certain assholes the satisfaction of me asking you

So I'll just waste more time, whatever

Danu

Benny Benassi - 'Satisfaction' (Official Video)

►

John Rennie

I think you underestimate the good nature of the banter hereabouts

Danu

^that

Or pretend to do so for unclear reasons

0celo7

16:26

No, I overestimate it.

John Rennie

Why do you mainland Europeans all like Eurodisco?

Danu

Do you think anyone actually likes that?

John Rennie

Someone buys it

Danu

Not me, or anyone in this room for that matter :P

(I hope)

0celo7

Let me listen to it

I'm pretty open minded

Not bad.

Danu

16:32

Obvious troll is obvious

0celo7

Whatever.

The sound effects are annoying, but the actual song isn't bad

John Rennie

Actually I'd say it was inoffensive enough. If it was on in the background I wouldn't run away screaming.

0celo7

@JohnRennie is there offensive eurodance

Hardcore?

Not a lot of people like that

Hellsystem - Devil Face

►

John Rennie

There is some cheesy Europop around. Very cheesy.

0celo7

I think that's pretty good, don't know exactly what it sounds like off the top of my head though

David Z

16:35

@JohnRennie that's a low bar

0celo7

Actually that song sucks

John Rennie

@DavidZ when you hear a single on the radio it can often sound quite good, but if you are foolish enough to sit down and try and listen to a whole album it rapidly sears you.

David Z

Sure, if the other songs on the album are bad

John Rennie

Singers like Taylor Swift and Katy Perry are like that. You catch the song on the radio and think it's quite catchy. But I defy you to listen to a whole Katy Perry album

I tried and failed.

0celo7

@JohnRennie I'd probably enjoy it

I can't do large amounts of heavy metal

The growling gets old really quickly

DanielSank

16:38

I have some data processing procedure which results in a bunch of random numbers. I have computed an analytic expression for what I expect the probability distribution of those numbers to be.

How do I check it?

I had a conversation over at Cross-Validated, but I get the feeling that guy was going into more detail than I need.

Surely someone here has done this. Astro people use probability a lot, right?

John Rennie

@0celo7 that's death metal. If you listen to some of the early Black Sabbath or Led Zeppelin albums I think you be surprised how unlike the current HM bands they are.

0celo7

Ok, death metal

I can enjoy one of those songs here and there

David Z

@DanielSank I've done that. I think I even asked a question about it. I've never found a fully satisfactory answer.

0celo7

I can listen to hardstyle/any hard dance for hours on end

DanielSank

@DavidZ Wow, really?

David Z

16:40

Well, I've had the same problem. I think so, anyway, based on how you're describing it

DanielSank

How about a broke-ass answer?

David Z

hm. Similar, not really the same I guess. See this:

3

Q: Is there a way to test a probabilistic prediction for an uncontrollable experiment?

This is something that I've wondered about mostly in connection with weather forecasting, but I would like to ask the general question because it comes up from time to time in other contexts. Suppose I have a process, denoted $F$, which, given some initial state $x$, produces an outcome $y$ from...

probability hypothesis-testing

0celo7

Why not just plot it a bunch of times and see if it fits?

ACuriousMind

@0celo7 Very scientific :P

DanielSank

Ho boy.

Shouldn't I be able to ask the question "does my numerical data set with a million trials reasonably well match the cdf I computed?"?

(I actually computed the pdf, but whatever)

David Z

16:44

Yeah, you should. I'm fairly sure there's a known way to do that, but I don't know my way around the applicable math.

DanielSank

Let me put it like this: I can plot a historgram of the data against the computed pdf and visually see that they match really well. What's the computer version of that?

David Z

It might be one of the things linked in my question.

Oh maybe this is the one to go for:

Pearson's chi-squared test

Pearson's chi-squared test (χ2) is a statistical test applied to sets of categorical data to evaluate how likely it is that any observed difference between the sets arose by chance. It is suitable for unpaired data from large samples. It is the most widely used of many chi-squared tests (e.g., Yates, likelihood ratio, portmanteau test in time series, etc.) – statistical procedures whose results are evaluated by reference to the chi-squared distribution. Its properties were first investigated by Karl Pearson in 1900. In contexts where it is important to improve a distinction between the test statistic...

> It tests a null hypothesis stating that the frequency distribution of certain events observed in a sample is consistent with a particular theoretical distribution.

DanielSank

What's with this phrase "null hypothesis"?

@DavidZ This looks relevant, thank you.

David Z

np

@DanielSank seriously?

DanielSank

@DavidZ ...yes?

David Z

16:48

you never know here

DanielSank

Am I being dumb?

David Z

I have something else that might help you

Danu

Lol

David Z

this is from my blog, back in the day:

ellipsix.net/blog/2013/11/…

Danu

@DanielSank I thought you were doing stuff that involved data

0celo7

16:48

@ACuriousMind I've given up being thorough

Danu

How have you not heard of the chi-squared test :P

0celo7

I can prove anything

David Z

I think somewhere in that post it explains about null hypotheses

0celo7

@DanielSank has never taken a stats class?

DanielSank

@Danu I never said I hadn't heard of a chi-squared test. Don't give me that bs attitude "how have you never heard of xyz...?". It's not helpful.

Danu

16:49

Me neither, but those concepts are hard to avoid in a physics BSc

@DanielSank Sorry if I offended you...

David Z

I kind of get it. I never had to learn anything about statistics as an undergrad.

From what I've heard, it's frequently not a required part of a physics education.

DanielSank

@DavidZ Everything I know about statistics I learned myself, totally outside of school.

0celo7

@Danu oh screw you

David Z

@0celo7 I don't think that was really necessary

0celo7

You're never sorry when you offend me

How about he show some decency then

DanielSank

16:58

@DavidZ reading the wiki article on Pearson chi-square, it's what I thought "chi-square" meant in the first place.

John Rennie

@0celo7 Danu isn't trying to offend anyone. Any offense you've taken is unintentional.

ACuriousMind

@0celo7 wtf are you talking about

DanielSank

I computed the "statistic" and got 0.0008

To be specific, I computed (1/N) \sum(O_i - E_i)^2 / E_i

David Z

@DanielSank yeah, I think Pearson's test is the "standard" chi-square test that always gets used for these sorts of things

DanielSank

Where O_i are the observed counts in bin i and E_i are the expected counts in that bin.

David Z

16:59

Well, if that's not right, I'm not the one to tell you. I'm not an expert on it or anything.

DanielSank

@DavidZ Right so isn't this somehow appropriate for the case that I expect the distribution of counts in a given bin to be Gaussian?

David Z

I would think so

Transcript for