@TerryBollinger No, but I almost got in trouble for having Linux on my calculator during an exam

@DanielSank Yep, lol

@DanielSank Though if you live near a nuclear power station now might be a good time to move house :-)

In my defence I was just using bc

@DanielSank There is a very good reason why we use Excel spreadsheets. Care to have a guess why?

@JohnRennie ಠ_ಠ

@JohnRennie Sure.

Ummmm, is it because that's what the customer is used to?

@DanielSank I'm curious too...

Or is it because they integrate with MS Office?

It's the only thing the customer knows how to use

Legacy support?

It's because if I do it in C++ it has to undergo long and expensive evaluation by the IT department and it's only allowed to be installed by the IT dept. The total cost is several times what I charge.

That is absolutely terrifying.

Ah

That is horrible

Please note all: Excel and visual basic are actually both very powerful languages, just proprietary!

Whether or not the language is "powerful" isn't really an issue in my mind.

So John Rennie has a pretty good strategy for results there.

VB.Net was pretty weird in my experience

never used VB6

@JohnRennie are you saying that by running code in Excel's VM you avoid rules about code review?

@TerryBollinger I have done massive calculations in Excel VB, as in tens of thousands of lines of code. And it runs amazingly fast.

is the chat session still going on?

@heather not formally, but it never really started formally to begin with

@DanielSank I don't avoid anything. I do what I'm paid to do.

@JohnRennie ~sigh~

You said that if you write the code in C++ it needs formal review.

I'm heading out folks

@JohnRennie yep! And I'm the guy who just said that you-know-who once tried to get me fired, so it' not like I'm rigidly pro-Microsoft... :)

@TerryBollinger It was a pleasure

@JohnRennie @DanielSank @TerryBollinger Get me a job

@DanielSank the code is extensively tested by two different teams of nuclear scientists and not sogned off until everyone is happy.

So what code review is avoided by using VB? I am confused.

You avoid the UI needing review?

@BernardoMeurer I tried. Remember?

The IT dept know nothing about chemical engineering and care less.

Ungrateful little...

Great session all! John, I heartily approve of your approach -- you focused on the data and getting it done, not abstract arguments about how.

@DanielSank Yeah but I was "too young"

@DanielSank What the IT dept charge (internal charge) 20 grand to do I have no idea

@TerryBollinger could you please not make this about whether or not someone "approves" of JR's approach? I'm trying to understand the situation, not pass judgement.

@JohnRennie You're not answering my question.

@DanielSank what question?

You said that using Excel/VB avoided some review process. Could you explain that?

@BernardoMeurer, have a good day

@DanielSank Yes. It avoids the IT dept wanting to do some checks that they consider important but no-one else does.

What those checks are i don't know but they have nothong to do with the science.

@JohnRennie So their rule is "review stand-alone executables, but not stuff that runs in a VM"?

@DanielSank sorry, I wasn't trying to argue, didn't see your thread! I just meant it as a genuine complement, coming from someone (me) who has some non-trivial background in software development strategies and issues (former IEEE Software assistant Editor-in-Chief),

@DanielSank I suspect this is a matter of what they don't know won't hurt them.

@TerryBollinger ::nods::

@JohnRennie Why can't they not know about a C++ program?

You're asking me what motivates the EDF IT department.

In other words, why is it easier for you to distribute an excel file with a VB backend than it is for you to distribute, say, a C++ program.

@JohnRennie No, I'm not.

I'm asking, at a mechanical level, why you are able to distribute excel files with VB code but not C++ code.

I am having a hard time imagining how that's technically possible.

Ah, OK, the end users can't run executables that haven't been properly installed. That's what they tell me - I don't know the details.

I suspect this is a Windows software group policy setting.

@JohnRennie Ok, so as I said before, you avoid this restriction by distributing code that runs in a VM.

In this case, the Excel/VB VM.

You know that everyone has Excel already installed, so you can target Excel and the user won't have to do anything special.

Well, the techies at EDF use Excel, and they create Excel spreadsheets all the time.

This is just another excel spreadsheet.

Yeah ok I think I get it.

Thank you for explaining.

@DanielSank C++ is actually much more difficult language than visual basic to review at high quality, both line-for-line and function-for-function. VB is also a hugely used language, more broadly than most of us from the C++ world (includes me) tend to realize. That makes more expert reviewers available at many location than for C++.

@TerryBollinger Ok but that seems not to be John's issue. It seems, to my understanding, that by using Excel he entirely avoids review altogether.

...for the Excel part.

@TerryBollinger I just wish that Excel VB did objects properly. It does have objects but they are clunkily implemented.

@JohnRennie yep!

@DanielSank review by the IT dept. There are still two levels of science review.

@DanielSank good point and not just for Excel! Do you realize how much of the Linux/Unix world skips review by using shell scripts and other non-compiles?

@DanielSank as it happens this last project was a rewrite and extension of an old VB6 app. During the rewrite I discovered a major bug in the calculation they've been using for the better part of twenty years. Now that's scary!

@TerryBollinger Not sure what you mean. If someone submits a shell script to my repo at work, you best believe that thing gets reviewed!

@JohnRennie Even for the Excel front end to the already-existing VB code?

@JohnRennie Is the code tested?

@DanielSank not in the sense that devs mean by testing, no. The functionality is tested by doing a suite of test calculations.

@JohnRennie That sounds like a real test to me!

Actually I'm beginning to think might have been a bit too candid in a public arena. Now might be a good time to STFU.

If you get the right answer on known cases, that sounds like a reasonable test suite.

@JohnRennie Hahahaha, ok.

@DanielSank that's good! I've just seen maaaany examples in the past (times change) where that was not the case.

@TerryBollinger We review all commits.

In fact, you can't merge until someone submits a review.

...and the tests have to pass.

@DanielSank ah, yes, I'm talking about older, less automated processes. The progress and level of automation in support of processes over the past decade especially has to me been one of the most amazing and interesting aspects of emerging SWE. I have found it particularly fascinating because certain aspects of it parse the problem in ways that make artificial intelligence assistance more plausible for insertion into processes.

@TerryBollinger Do go on.

@DanielSank argh, I'd have to dig for which tools in particular, but the drift is this: The more that computers can coordinate the overall development process, the better the memory of it and the less likely it is to drop things like humans do.

The key is to parse development into purely creative tasks that require as little context as possible to explain, with the computer filling in the context. Humans then are tasked with these highly focused creativity problems at which they are best, and at which computers (still) stink.

I believe I understand your meaning and I agree.

It's no good to have the humans doing deterministic tasks.

This to me is one of the most fascinating aspects of recent tools, and I think it will end up being an oddly tangential important contribution to the "how do we make a true AI?" question. A true AI, to me, is one that can program. That is a very, very hard AI problem, likely one of the hardest of all, far more difficult than GO or chess.

Must go! Talk to you later...

ciao

@BernardoMeurer, please be sure to put some paprika on your next Elmer's...

@DanielSank Cheers!

^ o_O ?

@TerryBollinger, do have a good day

Umm. So Hi everyone.

Just saw my Old C++ books :P

( And of course, got emotional )

@heather @DanielSank Which language is the most preferable and profitable to learn?

@SwapnilDas, Python 3, but I'm very biased, so.

@heather If I barely remember, is there something called concatenation in it?

@BernardoMeurer I may buy the quantum mechanics audio books. but impossible to buy all - there are 20 audio books, and each of them costs somewhere around $33 usd.

@SwapnilDas string concatenation?

@heather Yeah!

that's in a bunch of languages.

but python 3 is still the best =D

My old Code-Academy days ;P

@heather Ok! reasons?

1. very clear to read, reads like english

ok fine.

print "Hello World" vs

class HelloWorldApp {
    public static void main(String[] args) {
        System.out.println("Hello World!");
    }
}

in another language

for example

2. this goes with number 1, it's less syntax-dense

it is indentation not begin or end or parentheses or brackets everywhere.

I see.

(::glares at lisp::)

Finally home

3. Monty Python's Flying Circus

^yes, it has a great name

4. it is commonly used

@heather Why don't you do a PhD, you know a lot!

@Slereah Congo! Where had you been to?

@SwapnilDas, I would fail a PhD so miserably it wouldn't even be funny

anyway

Oh seriously? lol.

The family

@Slereah Nice.

Very dull, actually

@Slereah You, currently?

not anymore

5. it just has some absolutely awesome features

@Slereah Do some GR ;)

^yes, that was a "it's awesome because it's awesome" tautology

@heather Sounds cool!

Well I've been reading Kay's paper in the plane

6. you can do object oriented programming, or you can not

But currently I have to read up on wavefront sets

so it's very versatile

Otherwise it's not gonna be fun

@Slereah Full name of the scientist please?

Fuck I dunno

shall i go on?

It's a paper by Kay and Wald

Kay Mc. Mr. Kay

On QFT on compactly generated Cauchy horizons

@SwapnilDas I'm in middle school, how would I do well in a PhD program?

I like C++ ^_^

It's the paper on QFT divergences on the Cauchy horizon

@heather Extra talenteds are allowed everything :P

I am getting close to understanding what the fuck is happening

@Slereah Ooh. Maybe I'll be consulting you seriously after 10 years.

@SwapnilDas I'm not an extra talented, but beyond that, i'm unclear as to your meaning

Prodigy :P

nope, definitely not

oh, just thought of another extra nice thing about python: many great packages

ic. Ok bye, thanks for this :)

Got to have a sleep :)

yup, have a good night

@heather My favorite python package:

@rob, that has now taken the place of best xkcd ever =D

No, this is the best XKCD ever. All experimental scientists should read, learn and inwardly digest.

@JohnRennie @rob you're both wrong

this is the best xkcd:

@JohnRennie oh, I get it...though I'm not sure if the comic helped me =P

@AccidentalFourierTransform Note that we have no evidence whether the stick figure in the "wrong on the internet" cartoon is wearing clothes.

@SwapnilDas Python is rather popular and widely used in science.

C++ is also widely used.

Java is good if you are interested in working more generally as a programmer because it is widely used in business.

If you want to make websites, you need HTML and JavaScript.

@DanielSank, I'm doing something very wrong in my math over the bit-flip code.

But I don't know enough to find the error, would you mind looking at it?

@heather sure

What's up?

the circuit I'm using is off to the right

i'm trying to simulate the bit-flip code

because my dad challenged me to simulate it and prove that I had error-corrected properly

and so I'm trying to work it out mathematically

but I know I did something wrong somewhere because every single qubit is flipped from what it should be by the end

and I have no idea what I'm doing wrong.

@rob imgur.com/MQDAepB.png

Not sure what $$\left[ \begin{array}{cc} 0 & 1 \\ 1 & 0 \end{array} \right] \left[ \begin{array}{c} 0 \\ 1 \end{array} \right]$$ means.

The matrix on the left is a NOT gate, but what's with the vector?

the 0 state

i probably messed up a bunch of stuff

Normally, when we write matrices and vectors, we order the basis states $\{|0 \rangle, |1 \rangle \}$.

So, $|0 \rangle$ would be represented as $$\left[ \begin{array}{c} 1 \\ 0 \end{array} \right] \, . $$

darn

okay, i'll fix that

In other words, @heather, if you have a linear operator $T$, the $ij^\text{th}$ matrix element is $$T^e_{ij} = \langle e_i | T | e_j \rangle \, . $$

i think that makes sense

We actually proved that in Linear Algebra Happy Fun Time.

It's kinda obvious though. Think of acting a matrix on a basis vector, i.e. one with all zero entries except for a single 1 entry.

Suppose that 1 is in the jth slot.

Just do the matrix multiplication and you'll see that you get the jth column of the matrix.

oh, that does make sense

So the ith row of that column is element $T_{ij}$, and you got it by asking for the ith element of $T$ acting on the jth basis vector.

okay

brb

got to help my dad real quick

k

Awwww, Carrie Fisher died.

:(

okay, back

@DanielSank =(

Yeah

=( is right

=/

No spoilers!

oh gosh

sorry

No problem. Nobody else is around.

phew

::ahem::

Probably should delete the other one too.

dangit

@ACuriousMind You mods and your lurking.

darn

I lurked waaay before I was a mod :)

@ACuriousMind, you didn't see anything...

Whatever. I heard ACM were an AI. He can delete the knowledge if he doesn't want it.

true.

@DanielSank, so besides the notation mess up, what else is wrong?

@heather not sure, I didn't look beyond the bases being flipped.

What is the goal here, anyway?

to "prove" that in my simulation I error corrected the qubits

so the goal is that if there's a bit flip it will be undone in the end?

basically to prove that, yeah

and i've got to do it well, because it's a project from my dad.

Ok, well the first image you sent me is just some state preparation and then a bit flip error (or not), right?

right

I'm not seeing a simulation here.

see, that isn't the simulation, the diagram of the circuit i did for the simulation is off to the right

in those images I'm going through the math, how each gate is applied

if that makes any sense

i'm explaining this poorly, i know.

The first six gates do nothing but make the $|111\rangle$ state, right?

yes

going to go, but I have my phone on me.

okay

hmm, wait, i've got to have lost track of something somewhere

because replacing the last 5 gates (before measurement) with a toffoli gate

gives the right solution, i think

wait, no...

i don't know

I'm also puzzled by why there are measurements at the end.

the online platform i was using for the simulation required there be measurements or it wouldn't give any data.

which i guess makes some kind of sense.

hmm, i must've done something wrong because i think if the last gates were replaced with a toffoli gate, the first qubit would come out in the |1> state, which is right

but the second and third qubits would also come out in the |1> state, which still isn't right

Hey guys

@PhysicsGuy, hello

Got the blue foot hat.

It doesn't fit an octopus well.

Tophat and specs work much better.

@DanielSank I'd accept it from you, if that were possible...

Shame hat trading isn't a thing :(

Anyone here an astronomer?

@TerryBollinger Student's don't have money for Paprika :P

@ACuriousMind @DavidZ Any of you two around?

@BernardoMeurer what do you need?

@ACuriousMind pgfplots help

\pgfplotsset{
	axis lines=middle,
	axis line style={->},
	xmin=0,
	xticklabels=empty,
	xtick=\empty,
	ymin=-1.2,
	ymax=1.2,
	yticklabels=empty,
	ytick=\empty,
	samples=200,
	every axis plot post/.append style={very thick},
	clip=true,
}% end of common axis set
\begin{tikzpicture}
	\begin{axis}% 1. plot
		[
		domain=0:4*pi,
		xmax=pi,
		yscale=0.5,
        ylabel={$\psi_1(x)$},
		]
		\addplot[red] {abs(sin(deg(x+pi))*-1)};
	\end{axis}
	\begin{axis}% 2. plot
		[
		domain=0:4*pi,
		xmax=2*pi,
		yshift=-3cm,

I have never used pgfplots

This should generate a probability density graph for n=1,2,3

Dangit

How do you make your plots?

This is what I'm getting, but if you look on my plot commands I'm plotting the abs value!

@BernardoMeurer What plots?

@ACuriousMind Idk don't you ever need to do plots?

@BernardoMeurer not really, no

Dangit

Go do something that needs plots

learn it and then come help :P

@BernardoMeurer you like open source right? try gnuplot