@El'endiaStarman what about letting the users choose a subset of 3 of those 10 basis vecotors, and then displaying the projection to those 3?

I think I've already seen plots with up to 6 dimensions displayed by also including colours and sizes of symbols or similar things.

ps source

@El'endiaStarman are you trying to show data? If so, you can use more than spacial dimensions (color, for example)

@NathanMerrill The data itself is actually relatively unimportant; it's the relationships between data that's the focus of this.

@flawr That could be workable. It'd be a bit like Blender's views along axes though, which can miss out on some actually interesting structure that you might see from another angle.

This is this SO tag data put into 3D: i.imgur.com/afWGfQl.gifv

@El'endiaStarman are you able to programmatically detect some relationships? Like, there could be lots of false positives or false negatives, but choosing each set of 3 vectors out of 10: There's a lot of possibilities there?

some sort of heuristic to try and identify interesting stuff

yeah, 10!/3!= 600K

wait...that math is wrong

10 choose 3 is 120

there we go

Yeah, I was thinking of programmatically finding large cliques to determine the upper limit on dimensions, and then for interesting stuff....uh, I dunno.

also, is it just me, or are you guys also amazed that the math behind N choose M never comes out as a fraction?

like, 10 choose 5 is 10!/5!/5!

I still don't understand how that second 5! just happens to have all of its factors in the 10-6 range

I wish there was an option to say that I don't intentionally skip meals...

I was just about to say that

@NathanMerrill think of it the other way round: 6-10 are five consecutive numbers, so there must be one divisible by 5, one by 4, one by 3 etc.

I nearly always forget about lunch

@MartinEnder how do we know that it isn't 1 number that divides them all (and therefore, doesn't have enough factors to share?)

I'm not sure what you mean

("all" is a bit too strong, but you get my point)

maybe not :)

like, for example, 12 divides 6 and 4 and 3 and 2

@Feeds Because people clump together. [watches video]

but you can't do 12/6/4/3/2

but 12 is only one number.

So, I get that a factor of 5 must exist somewhere in the range 6-10

and similarly for every other factor

and so must a factor of 4 and a factor of 3 and so on

but how do I know they won't all land on the same number

(or at least, most of them)

well for a start, once you get to smaller divisors, there'll be several factors you can choose from.

okay, 5 is probably a bad example

let's take 6, where 4 isn't coprime with it

and say we've got 10*11*12*13*14*15

now both 6 and 4 land on the 12, but we can't divide by both

that's the issue, right?

exactly. And in this case, you can move the 2 to a different factor

yes

and I agree, that that always happens

but I have no idea why

its like magic

and yes, 2 is more common than other factors, so it's more likely to appear, but it won't always be a low factor like 2

@El'endiaStarman Okay, the answer is more like "people clump together as much as technology allows".

you can probably just break the argument down to prime factors. instead of saying, six consecutive numbers add at least one factor of 6, you can save they add at least four factors of two and two factors of three, which covers all of 2,3,4,6

@Feeds this video has one of my favorite economic concepts that I want to push a bit further

he was talking about pencils and paper. Person A can make Pencil in 30 minutes and Paper in 60 minutes, while B can make a Pen in 60 minutes and a Paper in 30 minutes

by trading a pen for a pencil, they both save 30 minutes

however, this can get even more extreme

let's say that Person B is really slow, and can only make a Pen in 4 hours and Paper in 2 hours

still, by trading, I've saved time. Person A makes 2 pencils in 60 minutes, and Person B makes 2 papers in 4 hours

they trade, and both save time

even though Person A is strictly better at both tasks, they both save time because the ratios are different

@MartinEnder that is a nice argument, I only ever convinced myself by deriving the formula for the binomial coefficient using the binomial expansion itself:)

@NathanMerrill But if person A makes two pencils and two papers, then that takes them 3 hours, which is less than the 4 hours it takes for person B to make two papers.

@El'endiaStarman but then he has no reason to give anything to person B.

so a trade won't happen, so person A will have wasted 2 hours on stuff he doesn't sell

If they use computers they don't have to make any pens or pencils.

lol

actually, at my work (in software development), I often look for a pen/pencil and can never find one

I'm not sure if that proves the obsolescence of pencils or proves the opposite

Actually I use a ton of paper, and prefer it for many tasks over any electronic devices.

I received two pads of paper when I started working. One's at home. I also have two "books" of post-its, a pencil, two pens, and eight dry erase markers (and a whiteboard).

I just miss the ability to link stuff or use crtl+F when working with paper :)

Augmented reality!

Shouldn't it be actually called diminished reality, as AR makes things less real?

lol

...I would quote Kirito but I don't know if you guys have seen (or liked) Sword Art Online.

@flawr since reality is the very definition of real, trying to make reality more real is like trying to make red redder. I'd expected "augmented red" to actually be less red than red

But augmented reality is not trying to make reality more real. It is adding things to reality.

So there are more real things in reality that we can perceive, not that reality becomes realer.

so, augmented reality is really augmented perception

If a tree falls in an augmented forest and no one is there to see it, does it make a sound?

did it fall?

<philosoraptor/>

TIL that material != materiel.

one is german, one is french? :D

@El'endiaStarman but by adding stuff you still lose other information

@El'endiaStarman let's say there is a blind man in the vicinity...

@flawr Clearly, blind people are incapable of hearing in general.

@NathanMerrill that is why they have hearing-ear dogs

The best part is the hover text

@NathanMerrill Hahahaha yes.

@NathanMerrill Had a random idea for KotH. I'm not really planning to write a controller for it, but it seemed like it might be your thing, so here goes: if there are N participants, every bot gets randomly assigned a unique strength from 1 to N, but doesn't know its strength. Now you alternate between a) each bot guesses its strength, and b) each bot fights another bot and gets to know what its opponent thinks how strong they are, and who won the fight (winner is determined by actual strength).

you do that for some time (e.g. a full round robin) and at the end the bot who guesses its own strength most accurately wins

@MartinEnder So, winning the fight is irrelevant?

aka, I could lose 1000 games in a row, and still win

...it's like a collaborative sorting algorithm

I totally might go with that theme: "Write me a self-sort!"

@NathanMerrill that's what inspired it

@NathanMerrill yes, the goal is to figure out how strong you are, you can't do anything about the actual outcomes of the fights

I really debating whether or not to tell the users how many steps are left

because if I do, then malicious bots are viable competitive

what?

how do you strikethrough on chat?

Triple hyphens I think.

there we go

I'm mostly worried that taking your winning percentage as a direct estimate of your strength might give a really good result, even if you ignore your opponents' estimates entirely.