Back to my previous point: GWT, Typescript, Brython, asm, and the list goes on.

am not disagreeing with any of this but a big part of web dev is "RAD". and not a lot of overthinking the language. have indeed seen a lot of crap. and some of it is in production, live and serving up tons of traffic on the web...

there is room for different povs.

there is no "one size fits all" or never will be...

indeed

js has serious cross platform compatibility issues for many years, some of it now ironed out, esp with advancing browser versions, some of it insulated by jquery etc.

@vzn That problem can only be fixed by the browser developers. Not sure how it's relevant to what we were discussing.

Oh I see what you mean.

jquery is a library, not a language, by the way.

ok. jquery is close to a language at this pt.

meant to say cross browser incompatibility but cross platform incompatibility is similar.

@vzn I knew what you meant.

believe it or not, jquery was originally built in analogy to SQL! its expanded way past that ofc.

Do whatever you want, of course, but I am telling you from experience that one should avoid using dynamic type-unsafe languages as much as possible.

@DanielSank web dev has very difficult injection attack problems that are language cross cutting. even built into original html design in a way.

They're fine when your project is teeny-weeny (and even at that point they force you to write unit tests for things a compiler would check for you), but the second your project becomes large enough to be interesting you find your attachment to dynamism is now a painful addiction.

@DanielSank think the debate/"controversy" about typed vs untyped languages is largely academic.

@vzn I don't understand what you mean. SQL injection attacks are an issue with SQL, not the web.

@vzn I don't know what debate and/or controversy you're talking about.

@DanielSank "injection" is a general attack pattern that is largely language crosscutting.

This isn't an academic issue. Large projects are almost never written in dynamic languages. It happens, but it's uncommon. That's just a practical fact of life.

@vzn That's true. HTML is not a programming language. It's type-setting, like TeX. Using it to do logic opens up all kinds of problems (because it comes down to string manipulation, which is hard to protect against injection).

@DanielSank dont know who you learned coding from, but this is an increasingly inaccurate statement.

@vzn Is windows, Linux, Mac OS, git, SVN, Chrome, Firefox, Opera, Safari, or any other large mission critical piece of software written in a dynamic language?

massive corps like google, facebook, etc are full of/ bulging at the seams with dynamic code.

@vzn Most of Googles core stuff is C++

The only official Google language which is dynamic is python.

Doesn't HTML5 allow programming or something

@DanielSank google is inventing hybrid stuff like "go". the distinction between dynamic and compiled is getting very blurry these days, to the point its again, maybe mostly academic.

I dunno

@vzn I'm dying to hear how go is even in the slightest degree "dynamic".

@Slereah ehhhhh, not in the sense I'm talking about... but yeah kinda sorta.

making something Turing complete isn't that difficult, usually

let me quote from this golang.org/doc/faq

A guy I know found out that by adding just one feature you could make wikipedia Turing complete

@Slereah Nope. TeX is Turing complete :)

If wave functions are complex numbers like $a + bi$, what do these numbers mean physically?like what are $a$ and $b$? Like for a given set of possible physical states, how does using complex numbers like a wave function allow me to define a super posistion between two states?

Happy holidays btw everybody

The superposition is the dumb sum of both

Physically they don't mean much

the wavefunction isn't a physical things

Only observables are

@vzn Yeah, but go is not dynamic.

If I say "Cars combine the power of steam engines with the compactness of horse drawn carriages", you wouldn't argue that my car runs on steam.

doesn't it

lol

it runs on gasoline steam

(As an aside, anyone reading this ought to take a look at Scala)

@Slereah Egh, ok fine.

@DanielSank not an expert on Go but afaik it supports either dynamic or compiled execution...?

"Go, also commonly referred to as golang, is a programming language developed at Google in 2007 by Robert Griesemer, Rob Pike, and Ken Thompson. Designed primarily for systems programming, it is a compiled, statically typed language in the tradition of C and C++, with garbage collection, various safety features and CSP-style concurrent programming features added."

It has a runtime environment that supports memory management but that's not the same thing as being dynamic.

@DanielSank Why Scala over say Haskell?

@alarge Well, Scala runs on the JVM, so you can deploy it easily on whatever platform you want.

Also, you can write Scala as regular procedural OOP code, so you don't have to use functional style when it doesn't make sense.

@alarge Haskell is neat too, of course, and some nice software has been written in it.

Scala also has the advantage that it can consume Java libs.

Not that I'm that versed with category theory, lambda calculus and all that jazz but I thought you could have stateful computation in FP through monads and such.

That means you have an enormous ecosystem at your fingertips.

@alarge They're all Turing complete, so you can do whatever you want, of course. It's a matter of style and convenience.

@vzn The big thing go offers (AFAICT) is coroutines.

I've also heard from several experienced folks that .NET is awesome. Since it's now essentially cross-platform I think it's quite interesting.

.net is aight

This is hilarious: Facebook uses some Haskell internally for tools. lex-pass is a tool for programmatically manipulating a PHP code base via Haskell.

Quote from Haskell's website.

@Slereah Used it much?

Somewhat

I like C# enough

It's a nice modern language

@Slereah How does C# do concurrence?

Is there a coroutine system or is it all callbacks?

Not a clue

C++ will have coroutines soon enough

@alarge ...how?

The C runtime is so simple, there's no stack virtualization. Are they going to use macros?

DS ok re Go as emphatically not interpreted/ dynamic. am surprised to hear your strong anti dynamic slant. dont you do a lot with python?

@vzn I didn't say I don't like python or dynamic languages.

Python is great.

I just have learned, from painful experience, that:

1) You wind up writing unit tests for everything, including a ton of stuff a compiler would check for you.

2) Run time errors make me much more sad than compiler errors.

3) Refactoring code without a type system to help you keep your interfaces right is a lot harder than refactoring code where the compiler tells you when you messed up.

@DanielSank Never used coroutines so don't know. Boost has had them for almost 10 years or so I think and my understanding is that they are expanding the C++ standard based on it.

@vzn If you ever write a piece of code that take a while to run, only to have it crash when the python interpreter hits an unknown variable etc, you'll know what I mean.

think large prjs are hard in any language and do require "conventions/ best practices/ good hygiene"

pycharm helps with all this, but it can only do so much.

@vzn Agreed.

some prjs overdo (automated?) unit tests etc

@vzn That doesn't address the unit tests, runtime crashes, refactoring issue though.

@vzn Ok. That's like saying "some people eat too much fiber and vitamins."

Yeah, maybe some do, but they're enormously outweighed by the projects which don't write good tests.

Lots of tests aren't the important thing. Good tests are.

I've seen people write "tests" that do nothing but check for syntax errors in python.

@DanielSank each prj is unique in some ways and some generalizations are for academics.

Think about that.

You know what's the real reason I like python?

1) Awesome standard library (especially numpy)

2) Simple "build" system. I was discouraged so many times in the early years of learning to program by not understanding how to actually build and deploy my code.

I'm looking at you, C++.

Python let's you just download the interpreter, type "pip install -r requirements.txt", and then "python my_program.py".

Of course, with java the situation is often similarly easy.

And once you figure out a sane build system for C++ it's not that bad, but figuring it out the first time can be horrible.

Hey, @vzn, have you tried Scala?

glad you apparently like coding at least as much as physics!

have heard some of scala, havent used it myself.

@vzn It's a hobby and a big part of my job.

@vzn Ok then you may find this really interesting.

Suppose I have a class which represents a stack of objects.

I can push new elements to the stack or I can pop the top element off of the stack.

With me so far?

ok

Suppose instances of this Stack class are immutable, which means they never change. In that case, when you push you actually get a new instance of Stack.

Now, it would be incredibly annoying if you had to write a Stack class for every single type of object. You don't want to have to separately program a Stack of Int, String, Fruit, etc.

Many languages let you avoid that by using generic types.

You write something like this:

class Stack[A]{ code goes here }

ok. Java covers this fine generally except for the primitives (added late to the language as a compromise for c++ devs)

Right, ok you're familiar with generics. Good.

Now suppose I have some code which expects a Stack[Fruit], but I give it a Stack[Apple]. Is this ok?

Yes or no?

in java either may happen. it depends on the generic implementation.

@vzn I'm not sure what you mean "either may happen".

one can have a class Stack<Fruit> or Stack<Object>, the latter being the more generic original implementation of some container classes.

Suppose I have a function func(Stack[Fruit] x) {...} and I pass in a Stack[Apple]. Is this ok?

@vzn I'm not asking about Stack[Object]. I'm asking if the code expects Stack[Fruit] is it ok to pass Stack[Apple]?

think that subclasses are ok in the generic system in your example. (havent tried it)

@vzn In Java it would work.

ie Apple subclass of Fruit & the compiler checks it. afaik. its not a common construction.

@vzn This is an incredibly common construction!

I might have a Stack[Person], and then a type Student which is a subtype of Person. Then I might have a Stack[Student] which I want to push onto a Stack[Person].

Think of it like this:

ok

Suppose some code is going to do my_fruit = my_stack.pop()

And suppose the type of my_fruit is Fruit.

Then, if my_stack is a Stack[Fruit], we're fine, right?

yes

If my_stack is a Stack[Apple], we're still fine, because pop() returns an Apple, which is a Fruit.

Do you agree?

right. some of this cant be checked at compile time. in java.

Runtime checks are bad! The point of my example is going to be that Scala does better!

Runtime checks mean possibility of runtime errors.

Well ok now you edited runtime --> compile time, so my comments look wrong.

@DanielSank lol division by zero is a runtime exception, so is EOF. etc

But you're right: some of the type checks happen at run time in Java.

sorry typo

Java has to do runtime checks because the following is legal in Java:

jlo was dancing on the tv :P

Do you see why that's a problem?

this is all great but to me OOP is old stuff, eg generics added almost a decade ago to java, and containers added ~1½ decade ago etc.

Ah, but here's the cool thing:

That bit of Java I just wrote is bad. The third line causes a run time error.

In Scala, you can check this statically.

See, in the Stack[Fruit] example, we said it's ok to get a Stack[Apple] because an Apple is a Fruit.

This is called covariance.

The Stack instance subtyping follows the type parameter (i.e. Fruit and Apple) subtyping.

However, consider the push method.

This fails, right?

In Scala, you can tell the compiler that you want the input to push to be contravariant.

That means that if I pass in a type which is more general than the Stack's type parameter, I get back a Stack of the more general type!

So the result of my_stack.push(banana) is a Stack[Fruit]!

This sounds too specific to be useful but the point is that the scala compiler understands type variance, so it eliminates the need for all the run time checks done in Java.

all OOP languages will continue to have runtime checks, it cant all be checked at compile time, this is nearly a provable fact. there is some adjustments around the edges that can be made.

do not think any of this "small adjustments around the edges" has any major impact on programmer productivity etc

how goes the qubits lately anyway?

(...sometimes different language features feel like different color drapes on a house to me...)

@vzn Does C++ have run time checks?

Or do you mean that the structures designed in any given language will have run time checks?

@vzn Well sure, but some colors make you want to vomit, which others give you a peaceful happy feeling.

@DanielSank "type safety/ checking" seems to inherently require some runtime checking, it seems to be "inescapable". analogous/ similar to division by zero exceptions. think this is possibly provable somehow. (it would require some major finesse/ good definitions.)

Interesting. What exactly is the underlying "reason" for divide by zero errors?

@DanielSank a sort of zen question along the lines of "what is the sound of one hand clapping"...?

No no.

Regarding 4 / 5 as 4.divide(5), the real issue is that the Float type is too general to be used for the divide method.

@DanielSank so sounds you have some ideas for a new language design? as they say in NYC "and good luck with that" :|

No no, I agree with you. This looks like a really general problem.

I have no idea how one could fix something like this.

@DanielSank do have some major appreciation/ admiration for language designers and the way that OOP theory made its way into very applied languages/ industry. its basically one of the biggest theory success stories in CS...

I guess there are cases where you want to define a function to take almost all instances of a certain type, but not all of them.

Weird.

but at the heart of it its an attempt to avoid defects which are inherent to coding, and will always be...

I don't know what you mean by that.

Divide by zero is tricky because it depends on the value of a thing.

I can write divide(float x, float y){return x/y}

That code is correct as long as the value of y isn't 0.

in some contexts (eg even java) divide by zero causes exception, in others it computes to NaN (Inf) and causes no exception, and in both cases that may not be what was intended etc (hence this being an interesting/ nontrivial test question that popped up for me recently)

I don't see a way to improve the code without a run time check.

a cohort at work is testing hackerrank hackerrank.com for our interviews.

@vzn Yeah, I suppose that's what exceptions (and NaN) are for...

it seems quite sophisticated, it has built in compilation, multiple languages/ dbs etc... was impressed by it, it will be an excellent diagnostic tool for us (versus humans).

Man, a lot of websites are starting to look like that now.

@vzn Uh, how do you avoid people Googling the results?

Are the interviews done live?

@DanielSank believe it or not its the honor system, and thats another tricky wrinkle (aka "can of worms") etc

the test says that "consulting std documentation is allowed" but these days, a lot of std documentation is on the web (eg java APIs esp), and its googled. and then its like, does SE count as "std documentation"? presumably not, but its a slippery slope.... the real dev will use these sources/ tools/ resources on his job, so why disallow them on tests? in fact would it be better to measure that also (ie google fu skills, SE comprehension)? etc

@vzn I would say that as long as you all understand the limits of your system, its ok.

all interview systems are limited, they are highly human oriented...! :|

I definitely think it's important to observe candidates in an environment where they're on their own though.

@vzn Yeah, we take great pains to work around those limitations. Interviewing is really hard.

in some ways the concern about software quality translates into hiring decisions, ie trying to find ppl that grasp it, and live/ code it, etc

for us the hackerrank test will likely be a pre-interview screening tool ("weeding out".) this has its own pros and cons.

@vzn It's good that you're aware of the limitations.

and actually asking someone to take a 90m-120m test just to possibly get an interview is not a minor request. but not uncommon these days.

@vzn Indeed.

I need to go to sleep.

Pleasure talking with you.

nice chatting happy holidays & hope the qubits get even cleaner. :)

(I once did a ten hour research sprint to get an interview)

@vzn Heh, we're working on it.

Ciao

similar to codejams eh? us too, more popular these days. (curse @#$% zuckerberg!) bye

Does the uncertainty principle mean that information cannot be measured or that it just doesn't exist (and hence can't be measured)?

@ChrisWhite lol no offense but you guys sound like models (or maybe their agents) at a fashion show. eg finding all the (copious/ glaring) flaws of the victoria secret models :P

however, the journals publish cutting edge areas and (agreed) there is more risk of mistake in that high-flying realm.

the list of broad crosscutting ideas introduced in science/ nature would be very long and include nearly all the major/ notable/ groundbreaking shifts of many fields.

re model criticism, have noticed a similar attitude in hollywood about the "blockbuster/ tentpole" movies. by fans/ critics etc... you guys probably think theres something wrong with the new star wars movie too, right? :P

@0celo7 maybe you have to, yes...I don't know exactly

@ACuriousMind in QFT, what exactly do "field operators" operate on? Do they operate on wave functions?

in principle once they accept the membership and you sign in, the discount should be automatic

@StanShunpike yes

@yuggib interesting, thanks

or more generally, on quantum states

Why do we need field operators?

Like what does this offer differently that ordinary quantum mechanics did/does not?

this is not different from ordinary quantum mechanics

the point is that the classical phase space associated with field theories is infinite dimensional

therefore the "canonical variables" are infinite dimensional as well; after quantization this yields the field operators (instead of usual pos/mom operators) that satisfy the canonical commutation relation

What does having an infinite dimensional phase space in the classical case have to do with field operators?

@ChrisWhite so a local positivist would claim that no measurement implies no information exists? Here I mean information as in like a particle has a position and momentum

that the field operators have the form that they have because of the infinite dimensionality of the classical space

otherwise, the operators would be (isomorphic to) the usual position/momentum operators

@StanShunpike AFAIK there is no consistent way to introduce Schrödinger-like wave functions in QFT. So, the answer to your question is, the field operators operate on quantum states.

The relation between wave functions and general quantum states is $\psi(x):=\langle x|\psi\rangle$, where $|\psi\rangle\in\mathcal H$ is a general quantum state, and $x_i$ is a complete orthonormal set of eigenstates of the position operator $\hat x$. If I remember correctly, there is no way to do this "wavefunction creation procedure" in a relativistic quantum theory

too late to edit.. I wanted to write that $|x\rangle$ is the position operator's eigenstate with eigenvalue $x$, so $\hat x|x\rangle=x|x\rangle$

you can introduce wavefunctions in QFT

it is not done a lot, though

What is the difference between a quantum state and a wave function?

one is a representation of the other

also

a quantum state doesn't vary in time, in the Heisenberg formalism

but a wavefunction in the Schrodinger formalism does

There are 3 "pictures" right? The third being the "interaction" picture. Does it matter which one we use in any given situation?

well more than 3

but those are the main one yes

More than 3?

and the third isn't the interaction picture

it's the path integral formalism

Wow, really? Hmmm interesting

ö.ö

the various formalisms are useful to prove certain things

Some things are easier to prove than others in a particular formalism

What is the interaction picture? Wikipedia listed it as the 3rd one

But it makes more sense to me

That the third is path integral formalism

Interaction picture is halfway between Schrodinger and Heisenberg

Right thats what I remember reading, like a hybrid

but there's a lot more formalisms

Phase space formalism, state density, lattice, stochastic, Bohm, etc

Algebraic QFT D:

Bohm is best

Omg

and the other axiomatic QFT, too

Wow

Thats amazing

I had no idea

oh and quantum logic, but that one has been given up on IIRC

So what kind of physics do you work with?

Currently none :V

Job market is not very nice for physicists

*in France

All the world is France

hon hon

But really QFT mostly just uses Heisenberg and path integral

that's about it

yes

only path here

Even Schrodinger is pretty rare

Except for some pretty specific applications

usually people use a weird mixture of the two

Because working entirely in one or the other is a pain

Proving unitarity with path integral isn't easy, and neither is proving the Schwinger Dyson equation in Heisenberg

When talking about frame bundles, Nakahara says "In general relativity, the right action corresponds to the local Lorentz transformation while the left action corresponds to the general coordinate transformation." Is this just a convention, or is there a deeper meaning to this?

@yuggib can you please check if they put on the discount when you hit "checkout" and they ask for credit card info

@Bass Where does he say this? I remember it but didn't give it much thought.

@Danu In your town btw.

@Bass I think it's because in the vielbein formalism we have the frame thingie $e^\mu_{\alpha}$ where $\mu$ is a coordinate index and $\alpha$ a Lorentz index. So, viewing this as a matrix, multiplication on one side is a coordinate trafo and on the other side a Lorentz trafo. So it's pretty much arbitrary. It depends on whether one writes $e^\mu{}_\alpha$ or $e_\alpha{}^\mu$.

@Slereah what? Schroedinger, Heisenberg and Dirac (interaction)

do you even first semester QM?

@StanShunpike One calls the QFT analog "wavefunctionals", and they're not really useful in most cases, so just say they act on the quantum states in the Hilbert space of states (which is just the Fock space for free fields, for instance).

@ACuriousMind Suppose I have some cont. function $f:[a,b]\to\mathbb{R}$. Suppose I also have $f(t_i)=g(t_i)$ where $g$ is also cont. Is the set of all such $t_i$ closed? Why? (Context: proof of Hopf-Rinow.)

@0celo7 Yes, it is closed. If $f,g$ are continuous, then so is $f-g$, and you're just asking whether $(f-g)^{-1}(0)$ is closed.

@0celo7: define $h = f - g$. what is $h^{-1}(0)$?

-_-

@Huy Don't know.

@ACuriousMind Well...I don't know how to answer that either.

what do you know about continuous functions

They're fun for the whole family

then you should ask your family

Preimage of closed/open sets is closed/open.

ok

so again what is $h^{-1}(0)$

@0celo7 So...you're really not seeing how that can be applied here? :P

Preimage of 0 under $f-g$

what is the "preimage" of $f-g$

I learned that in linear algebra...it's something

can you write down $h^{-1}(0)$ in terms of a set

@ACuriousMind Do you guys think I just make up problems

yes, aka trol

@Huy Hmm, I don't think so.

you should try

it's the kernel of $f-g$...

@Huy I'm scared you actually think I just troll all day long

you use the expression kernel for homomorphisms usually

write down the set

@Huy hmm, what's that?

usually something satisfying something like $f(x+y) = f(x) + f(y)$

no, not that

trol

the expression kernel thing

@0celo7 I'm not sure what step you think is missing. You did not object when I told you that you're just asking whether $(f-g)^{-1}(0)$ is closed. Then you told me you know that the preimage of closed sets is closed. So...what is your problem?

@ACuriousMind Lol, 0 is closed :P

That's the step I was missing, thanks

@ACuriousMind: if he can't write down the preimage explicitly maybe he doesn't know what the preimage of $0$ is

lol

@Huy $f^{-1}(A)=\{x\in X|f(x)\in A\}, f:X\to Y, A\subset Y$ IIRC

yeah, so it shouldn't have been very difficult to write it down for $h = f-g$

so what do you want me to write down explicitly?

and here $A = \{0\}$

no, I have no clue how to do that

idk man, you seem like you're just trolling really

@ACuriousMind Can you please comment?

@Huy please write down the answer...

surely you're capable of doing that too

@Huy no...what's the answer

please stop trolling

...why do you troll me like this

@0celo7 Conical singularities are defined to be counterexamples, essentially.

@ACuriousMind So a conical singularity has a well-behaved Kretschman scalar?

I thought a conical thing isn't even a part of the manifold.

@0celo7 "A conical singularity occurs when there is a point where the limit of every diffeomorphism invariant quantity is finite"

And indeed, if you just take the usual cone and equip it with the "standard" metric as a subspace of $\mathbb{R}^3$, you'll find you cannot take the tip into the manifold as such (for one, the tangent space at the tip doesn't exist, so you can't transport objects with indices to the tip), but every limit $r\to 0$ for any invariant exists.

@ACuriousMind Yes, but is that a singularity in the GR sense?

@0celo7 Ehhhh...what is a singularity in the GR sense?

Ah, Wald has the conical singularity as a counterexample.

@ACuriousMind Reading Wald to remind myself :)

@ACuriousMind I think it's just a place where geodesics end.

So...a geodesic cemetery?

Great, there are spacetimes which are timelike complete but not null or spacelike

@ACuriousMind Yes.

You should make that the tag wiki for "singularities" if we have a tag for that

Lol, Wald chapter 9.1 "What is a Singularity?" pretty much ends with ¯_(ツ)_/¯

Ah, screw GR, time to continue the Hopf-Rinow proof

@0celo7 section 9.4.2, in the paragraph immediately before the spin bundle dxample

So it is just a convention

lol watching some Goldfinger replays @0celo7

you seen the movie?

@ACuriousMind Do conical singularities occur on manifolds, though

@Slereah That's a strange question ;) Neither a conical nor a "proper" singularity is formally a point on the manifold.

@Slereah: What do you mean by that?

Aw

well conifolds are defined specifically for manifolds that are locally cone-like

So I am wondering if they are well defined on usual manifolds

Depends what you mean by cone. You could be looking to write down a singular metric on a manifold, where it's smooth except at certain points where it has a local form as the cone (ie looks like the cones you're thinking of sitting inside Euclidean space).

Conifolds allow you to have non-manifold points where the space is locally homeomorphic to the cone on some other space.

The former is a singularity in the metric; the latter in the topology.

Is there a link between the two

guys can someone please tell me why my answer is wrong. I deleted it here.

I guess one could write down a mutual generalization of the two, some sort of Riemannian conifold. I am not really interested enough in any of these to have anything interesting to say beyond the definitions.

@0537 Uh, if $\nabla\cdot\vec u = \nabla^2\phi$, then $\vec u$ is not divergence free unless $\nabla^2\phi$ vanishes.

yeah.

so if the left hand side vanishes and the right hand side vanishes its both.

that's what i meant.

@ACuriousMind is it right if i explicitly mention that?

@0537 Yes. I think it is much clearer to say "The gradient $\vec u = \nabla\phi$ of a harmonic function $\nabla^2\phi = 0$ is both-divergence free and irrotational."*

@ACuriousMind can you take a look at this...

any issues?

@0537 1. If you just correct your answer, or expand it, you should not post it anew, but edit the old one. 2. You're stating this very weirdly. You say a vector field for which both sides of $\nabla\cdot\vec u = \nabla^2 \phi$ vanish is irrotational, but that is only true if $\vec u = \nabla \phi$. If $\vec u =\nabla\phi + \nabla\times \vec A$, then it still fulfills that equation, but is not irrotational.

Also, what is a "harmonic vector potential"?

One calls $\phi$ with $\nabla^2 \phi = 0$ harmonic. That's not a vector potential.

@ACuriousMind I see, you're right i'll fix it.

Greetings

"The Analytical Rule might be considered a distant relation – as a skyscraper is to a shack – of that kindergarten toy, the logarithmic Slide Rule."

Asimov you are too old

@Bass Unless someone says otherwise, yes.