lol I think someone keeps going back and upvoting and answering all my questions :)

alephalpha does often just browse older challenges to resurface and answer

CMM: faq-proposed is a tag that anyone can edit in, but typically indicates that a post should at least be considered for the FAQ. Should we have a policy about when it is appropriate for users to edit in/what the appropriate usage should be?

time to make a proposed faq post about that tag

Tagged with faq-proposed-tag, of course :P

maybe we should add a tag to propose that a post should be faq-proposed

@hyper-neutrino Open a tag proposal for that.

faq-proposed-proposed-tag-proposal

beautiful

@hyper-neutrino is there a similarly simple solution for ax + by + cz == 0 ?

[z, 0, -x], [0, z, -y]

Wow. Thank you

i mean all equations of this form will be trivial

because there is only one equation and thus only one bound variable and the rest can be free

Is it only this simple because it is homogeneous? What happens if you have == 14 instead?

@Anush then that becomes more complicated i believe

Phew :)

How about if it is homogeneous but want the coefficients that give the shortest vector?

well you can just divide each vector by its elements' GCD

but that might not be optimal so idrk

Maybe that’s a more interesting question

you'd have to define that more formally; i'm not exactly sure what you mean by "coefficients that give the shortest vector" when there are multiple vectors in the solution basis

I am thinking of the coefficients a, b,c as forming a vector

So, minimize sqrt(a^2 + b^2 + c^2)?

Yes

that’s it exactly

but aren't there infinitely many solutions a,b,c

or do you just mean find the one single solution closest to the origin

and disregard the "return the general solution" idea?

@hyper-neutrino yes it’s a different question but you also need to specify that none of a,b,c is 0

This is a linear transform. So this is going to be equivalent to finding the point on preimage of a point closest to the origin.

i mean since 0,0,0 is a solution wouldn't you just end up trying to return as precise of an answer that isn't 0,0,0 but right next to it

like just set n_0,n_1,n_2 to like 10^-10 or something silly

or do you mean for integer a,b,c

You might want to say that the rank of the matrix is 1 because otherwise you won't have any solutions.

(Unless the value is zero in which case the every number is a solution)

@hyper-neutrino everything has to be an integer

@Anush then it's probably golfiest to just brute force lol

@hyper-neutrino yes that’s right,

@hyper-neutrino :) I want a method that finishes before the dinosaurs return

Yeah. You could probably come up with a convoluted module solution, but brute force is going to be way way better.

@GrainGhost only for small inputs I guess

what does LLL give you?

I mean better in terms of speed.

LLL is like hitting this problem with a jackhammer.

:) but a fast one!

The real solutions are parameterized by a plane and it's not hard to find the plane with a little linear algebra.

Ngl, I use brute force for my homework quite aa bit

If my coefficients are in the millions brute force seems a non starter

For what?

Is this supposed to be a code-golf challenge?

@GrainGhost just a challenge :) could be code-challenge if we decide that would be better

@cairdcoinheringaahing what sort of homework?

And your cooefficients aren't the biggest tax on run time, that's going to be the value on the right hand side.

@Anush Probability mainly

the main time i use scripts/programs for my homework is to run simulations to see if my answers for probability questions are correct (or look correct)

oh hey, ninja'd :p

for some reason, we do combinatorics in my probability unit

@cairdcoinheringaahing sounds cool

But the point is: If you want a fast answer you should start with linear algebra not with LLL. You can get quite far with linear algebra.

what's LLL?

"Lenstra–Lenstra–Lovász lattice basis reduction algorithm"

The name alone should convey how much of a nuclear option it is.

@Anush This sounds like a Jurassic Park kind of scenario. "We need to finish these calculations before the velociraptors come back!"

Yeah, that's... something :P

It is my opinion that it should be abreviated LLLLBRA for maximum Ls.

Also, the SO election ended with Stephen Rauch and Zoe being elected

@cairdcoinheringaahing Makes sense to me--combinatorics is pretty useful for probability, isn't it?

@cairdcoinheringaahing Combinatorics >>> stats

Yeah, its definitely linked. Just feels. Idk, weird?

@user Agreed. Stats (and probability) are evil

Do you get to come up with crazy scenarios where you offer your friends a choice between poison and ambrosia?

To be fair, my university had a Combinatorics class and a separate Prob/Stats class

The way statistics is taught is always very unmathematical. It's probably because statistics is a useful subject and the actual math around it can be very complex, but I find it kind of a shame.

We have an Advanced Combinatorics unit for second year I believe

@GrainGhost I don't know if it's actually the case, but it felt to me like you just try a bunch of different ways to do stuff and see which works best (e.g. when determining outliers, you almost arbitrarily pick a number to be your threshold)

I'm okay with combinatorics and probability, though I don't like how there are often multiple ways of approaching the same problem and it's hard to know that I got the right answer. Statistics? Yeah, not my cup of tea.

@GrainGhost I absolutely hate stats. I love maths for the fact that, of all the things in this world, maths is probably the most objective. If someone can prove that some statement P is true, there's no disputing that, beyond disputing the absolute core of mathematics. Stats however contradicts this. You can make any set of numbers come to any conclusion you like with stats, which violates the beauty of maths for me

@user I'm not sure what you are saying.

Bayes' law is pretty neat, though.

If you have to say "This statement is true, to a certain degree of uncertainty", that statement isn't true

@GrainGhost Like, you could say any number farther than 1.5×IQR is an outlier, but why 1.5?

@cairdcoinheringaahing That isn't too bad, and you can't always have absolutes in the real world

@DLosc Yeah, some of the axiomatic parts of probability are interesting, it's just the applications of them that I object to

@DLosc It's not even a law smh.

@user F*ck the real world, I have integrals to solve

@cairdcoinheringaahing It helps that I first encountered Bayes not in Prob/Stats but in Artificial Intelligence, one of my favorite classes.

@GrainGhost *theorem

"When will we ever use this in real life?" "Shut up man, I want to apply this interesting technique to solve the problem, who cares about 'real life'?"

Spoken like a true esolanger

Why bother with real life, when Jelly exists? :P

@DLosc ... It's not a theorem either ...

It's axiomatic.

@GrainGhost Then you have some Wikipedia editing to do. :P

It can be proven through the three axioms of probability measures, so it is a theorem, no?

@hyper-neutrino @Anush I studied the problem and general mathematical solutions yesterday, and I'm inclined to believe that HN's method (even with division by GCD) doesn't give fully general solutions

wait, why not?

oh, do the multipliers n_0, n_1, etc. have to also be integers

Yes, that's one

You define the concept of conditional probability, show it satisfies the axioms of probability measures, then show that Bayes' theorem holds

so like I need to find a set of vectors such that every integer solution is an integer linear combination of the integer vectors?

Yes, exactly

hm. yeah that's not at all clear in the question

@emanresuA that's what currently happens

Can you have an uncountably infinite partition of a sample space?

I'm guessing yes, but the concept seems... odd to me

@cairdcoinheringaahing Well, in actuallity you can choose your axioms how you want. Anything that's an axiom by one definition may be a theorem in another equivalent formulation. Bayes' law is sometimes an axiom sometimes a extremely simple result of other axioms, either way I think it's pretty silly to call it a theorem. Unless we are defining things in a rather contrived way (which sometimes happens).

Historically Bayes' law though is older than any sort of axiomatic treatment of statistics.

This is what happens when you bring up Bayes' law around me. I should come with a warning label about this.

@cairdcoinheringaahing ... Huh. I guess you can, if your sample space is uncountable (e.g. the real line or an interval on it). I think a partition of the real line can be defined as (or equivalent to) a set of real numbers representing the borders, in which case there doesn't seem to be a reason why that set has to be a countable set.

@DLosc Yeah, I was imagining a partition where each element in the partition is equal to a single real number in [0, 1]. That'd be uncountably infinite, and a disjoint partition, but I can't think of a "scenario" where that'd be the actual sample space

Probably depends on exactly what is meant by "parition", when we talk about paritions for e.g. integral definitions I'm pretty sure it's understood to be locally finite. Which sort of rules out uncountable.

I'm thinking in terms of the partition theorem

Ah yeah there it just means some class of disjoint covering subsets, so yeah if your set is uncountable you can have that by just putting each point in it's own subset.

Otherwise you are going to have trouble even evaluating the summation there.

I mean there's already some issues with having to evaluate countable sums.

@hyper-neutrino That doesn't make much sense. It could be considered unreasonable for PyGamer to not study humans, in which case it could also be unreasonable for you not to study particles. Besides, particles are very different from humans, and society's expectations for the two differ a lot. It's very much a double standard and one of the many things wrong with our society

Societal expectations for particles are pretty tough...I hear neutrons are under so much societal pressure to stay neutral all the time they can just randomly tear themselves apart.

And there's always the threat of antiparticles deciding to annihilate you

But think of it from the antiparticle's perspective. It's probably just as scared of you as you are of it :P

At least protons are always around to keep you feeling positive

And positrons, which are optimistic to the point of killing those negative electrons on sight

@RedwolfPrograms Yes, fortunately those never get older

@RedwolfPrograms But they're the enemy! Electrons may be negative but they can't help it, and they're one of us. Positrons are anti-everything we stand for

That's only looking at it from one angle though!

I'm not a particle physics, but I'm pretty positrons and protons tend to be pretty nice to each other

Anti-protons, however, are truly despicable

You're pretty positrons? :P

I sure am

@Wezlprogramsredwolf Saturday Morning Breakfast Cereal

But I'm not a physics of any sort

@RedwolfPrograms Oh dear, way I ask why? I thought they were just plain quarky

@user looks like a rip-off of super mario bros crossover :P

There's a new comic every day, so much bigger rabbit hole than xkcd (and lots of nerdy stuff for people interested in the particle war)

Well that's the fastest tag wiki approval I've ever seen

I guess whoever's in charge of slowing SE down took a day off

This can't be optimal can it? I've never seen JS beat J by 40 bytes before...

> I think there's a way to do this tacitly, but it'd probably become unreadable

> it'd probably become unreadable

> but

@RedwolfPrograms Yeah, it's got a for loop and stuff, doesn't look too J-y

A lot of older APL answers are also very, very long ¯\_(ツ)_/¯

The task itself isn't a great fit for APL family (and I guess most golflangs either)

It uses many variables and there's so little room for parallelization or anything

Oh, it's LYAL day again

What's the lang this time?

TIL you can't eat a grue

Your hand passes right through it, as if it was never there

do you have proof or have you just not tried enough times?

Anyway, gtg o/

Looks like the top voted are BQN, Underload and ///

Thoughts?

BQN was last time

I guess I should have deleted

ah ok. Underload, /// and Stack Cats then

@Wezlprogramsredwolf Nah, just edit in a note that it was already a LYAL

See the ><> or Cascade ones for example

CMP: For today's LYAL: Underload, ///, or Stack Cats?

I do have a fondness for Stack Cats, I won't lie

I like Stack Cats too, but I have a lot of work to do today so I actually want it next time rather than today