« first day (244 days earlier)      last day (155 days later) » 

14:13
@SimplyBeautifulArt Here I am
I have a very basic understanding of what ordinals are from a vsauce / Numberphile (I don't remember) video.
So our first thing is ∅, the empty set.
∅ = {}.
@Mr.Xcoder That's cool.
Ok, Ik about that
Next, we have 0 = {∅}
This is how we define 0.
Wait. So we define 0 as {∅}
14:14
go on.
And from here on out, we define x = {y : y<x} (set of all numbers less than x)
For example, 1 = {0}
2 = {0,1}
etc.
set of all numbers less than x - Maybe non-negative integers.
"numbers" refers to "ordinals", which are always greater than or equal to 0.
Non-negative.
@SimplyBeautifulArt ok. got it :D
14:16
Now, we can simplify this a bit
A set X is equal to x if x is the smallest number greater than all numbers in X.
So $\{1,2,3\}=4$
x = min{y : y > z, z ∈ X}
@Mr.Xcoder Yup
@Mr.Xcoder Should include 0.
But you can also simplify this down to 4 = {3}
@user202729 Unnecessary.
14:18
So we can do stuff like 100 = {99}
@SimplyBeautifulArt Ok, I was thinking about the previous version of your message.
And then we have infinite ordinals.
ω = {0, 1, 2, 3, ...}
... and we need extensions to ordinals? :P
ω = smallest ordinal greater than any natural number.
Yes, I knew that from that video (excuse the pun)
14:20
From there, we can go higher.
ω+1 = {ω}
ω+2 = {ω+1}
And we have $ω, ω+1, ω+2, ω+2, ω+3,...$
ω2 = {ω, ω+1, ω+2, ω+3, ...}
$ω^2$?
No $ω_2$ I think.
ω^2 = {0, ω, ω2, ω3, ω4, ω5, ...}
@Mr.Xcoder ω_1 and greater are not allowed for our purposes...
14:21
@SimplyBeautifulArt Then what is $ω2$?
@Mr.Xcoder ω2 = ω+ω
$⍵ \times 2$.
Smallest ordinal greater than ω + any natural number
@user202729 Dot products please.
:P Small nitpicks
14:22
@SimplyBeautifulArt Why?
@user202729 Eh, it's just how ordinals are done.
Anyways, we shall now translate these ordinals into finite numbers.
Oh I got it. I wonder what $w_n$ is though (now that I (mis)mentioned that)
So ⍵·2? (Funnily Enlist keyboard has both and ·)
@Mr.Xcoder ω_n is the nth ordinal with no bijection to lower ordinals... I can explain that later, but its not important here.
Ok. move on
14:24
The α-th infinite initial ordinal (from Wikipedia)
I remember that too
H(0,n) = n
H({x},n) = H(x,n+1)
H({x_0, x_1, x_2, x_3, ...},n) = H(x_n,n)
But loosely
These are the rules to the Hardy Hierarchy.
@SimplyBeautifulArt Wow let me process that :D
14:25
It's not so terrible.
The x1 is x·1?
No
Its a sequence
I clarified it.
So $x_i$ is a sequence where $i$ is ... natural number? Or ordinal number?
i is a natural number.
@SimplyBeautifulArt $\{x\}$ was equal to $x+1$ wasn't it?
14:26
In any case, we always take x_n, where n is a natural number.
So it wouldn't really matter much.
@Mr.Xcoder Indeed :-)
Then how is $H(\{x\},n) = H(x,n+1)$. Do we redefine the above?
The former one is $H(\{x\},n)$.
@Mr.Xcoder What do you mean? It is the definition of H.
Well, probably because of LaTeX strip off {}'s.
For example,
H(ω+1,2)
= H({ω},2)
= H(ω,3)
= H({0, 1, 2, 3, ...},3)
= H(3,3)
= H({2},3)
= H(2,4)
= H({1},4)
= H(1,5)
= H({0},5)
= H(0,6)
= 6
14:28
I don't understand the transition between $H(\{0, 1, 2, 3, ...\},3) $ and $H(3, 3)$
Index-of.
By the third rule.
Oh got it
Wikipedia has the same definition too. (some people somehow think that Wikipedia is hard to understand)
@SimplyBeautifulArt Ok.
Simple enough?
14:29
Yes.
Now we can get into some syntax.
{x} = x+1
Let X = {x_0, x_1, x_2, ...}
Let X[n] = x_n.
ω[n] = n
(a+b)[n] = a+(b[n])
(a∙b)[n] = a∙(b[n])
(a^b)[n] = a^(b[n])
How do you define + and ·? The same as in normal number?
And the ^ is exponentiation?
a∙(b+1) = a∙b+a
a^(b+1) = (a^b)∙a
Yes, they are the ordinal forms of addition, multiplication, and exponentiation.
So, now we can simplify our notation a little bit.
H(0,n) = n
H(x+1,n) = H(x,n+1)
H(x,n) = H(x[n],n)
If $x$ has finite element you use the second definition, and if $x$ has infinite element you use the third definition?
@user202729 yes.
14:34
I got this too
Also, ω[n] = n
3 mins ago, by Simply Beautiful Art
ω[n] = n
Yes
Same example as before using new syntax:
H(ω+1,2)
= H(ω,3)
= H(ω[3],3)
= H(3,3)
= H(2,4)
= H(1,5)
= H(0,6)
= 6
@Mr.Xcoder Oh lol
Are you using 1-based or 0-based?
@user202729 Hm?
14:35
Because $⍵$ is defined as $\{0,1,2,3,...\}$
@user202729 Doesn't matter. $x=x$ regardless of indexing
Apparently 0-based.
Oh, then yes, 0-based.
Now attempt to calculate H(ω^ω,2)
Wait a minute
You define $x+1$ as $\{x\}$
*::60 second counter starts::*
14:37
Give us some time. I'll post in a tio link for spoilers.
@user202729 yes
@SimplyBeautifulArt Please make that 300 seconds :D
so infinite values that is not a successor of any value have set size = infinite, while infinite value that is a successor of other value has set size = 1?
(although that finite set contains other infinte set)
@Mr.Xcoder *::adds on 270 seconds::*
@user202729 yeah.
@user202729 and yeah
Wait I forgot how $ω^ω$ is defined >_>
14:39
(a^b)[n] = a^(b[n]) if b is infinite set.
Ok thanks.
@user202729 Nope, definitely bigger than H(ω+1,2) = 6.
Also please spoiler first.
>! doesn't works here?
14:40
No idea how I should do a spoiler. You want TIO link?
I don't like TIO either.
>! Spoiler test
@SimplyBeautifulArt Yes.
So basically the result is base64-encoded and deflated.
^ Don't click unless you wanna check ur answer
14:42
@SimplyBeautifulArt This is what I have so far.
Weird spoiler method:
14:43
Nah, I like TIO better x'D
But I can't base64 encode and deflate in my head. This one is definitely easier.
Yay I was correct.
@Mr.Xcoder Use the syntactic approach plz.
12 mins ago, by Simply Beautiful Art
a∙(b+1) = a∙b+a
a^(b+1) = (a^b)∙a
@SimplyBeautifulArt What do you mean?
14:45
Simplify using that.
Then apply the other rules.
@user202729 There is one drawback: the content must be sufficiently short.
@SimplyBeautifulArt Better?
Also I'll have to go in 5'.
Also put = twice is not a good idea.
Big dot and not small dot ·?
14:48
Aww but I'll have to go and don't have time to fix. Rushing for now
@Mr.Xcoder Feel free to come back anytime.
@user202729 I don't have the small dot on my keyboard.
so...
Go make another keyboard (layout).
@user202729 @Mr.Xcoder by the time I finish, you guys will be blown away by the size of TREE(3).
:-)
Btw, once you get this down, you can approximately encode things like the Ackermann function @Mr.Xcoder
Hello @EriktheOutgolfer
@SimplyBeautifulArt Here we go \o/
not sure if I used proper notation. gtg o/
and cya
@user202729 shall we continue?
And do you happen to know the Ackermann function?
14:51
Yes.
H(ω^k,n) ≈ Ack(k,n)
So these numbers get very big, very fast.
Graham's number ≈ H(ω^(ω+1),64)
Conway's chained arrow notation goes up to about H(ω^(ω^2),n) I think.
And the Goodstein sequence grows so fast, we can't express them using H(a,n) yet.
To do this, we need a new number.
ε_0 = {1, ω, ω^ω, ω^ω^ω, ω^ω^ω^ω, ...}
(Note that a^b^c = a^(b^c))
Under this, Goodstein(2^2^2^...n times) ≈ H(ε_0,n)
Feel free to try and calculate something like H(ε_0,3)
(too big for me to post final result)
All good so far? @user202729
And then we can do nightmarish things like H(ε_0^ε_0,n)
While we're at it, we may as well make
ε_1 = {1, ε_0, ε_0^ε_0, ε_0^ε_0^ε_0, ...}
ε_2 = {1, ε_1, ε_1^ε_1, ε_1^ε_1^ε_1, ...}
etc.
@SimplyBeautifulArt So I'm trying to calculate this, but there is an issue: How can I detect whether a set is finite or not?
(in other word, should I apply rule 2 or 3?)
@user202729 As far as syntax goes, check if there's a '+1' at the end.
ε_ω = {ε_0, ε_1, ε_2, ...}
ε_(ω+1) = {1, ε_ω, ε_ω^ε_ω, ε_ω^ε_ω^ε_ω, ...}
@user202729 Would you like me to post my H(ε_0,3)? Spoiler/no-spoiler.
15:06
But you said that it is too large?
Yeah. I mean my expansion/work
(skipping a lot of steps)
I manage to reduce it to like $H(⍵^{⍵^2·2+⍵·2+2}·3,3)$
but it becomes too tedious so I give up.
H(ε_0,3)
= H(ω^ω^ω,3)
= H(ω^ω^3,3)
= H(ω^((ω^2)∙3),3)
= H(ω^((ω^2)∙2+ω∙3),3)
= H(ω^((ω^2)∙2+ω∙2+3),3)
= H((ω^((ω^2)∙2+ω∙2+2))∙3,3)
= H((ω^((ω^2)∙2+ω∙2+2))∙2+(ω^((ω^2)∙2+ω∙2+1))∙3,3)
= H((ω^((ω^2)∙2+ω∙2+2))∙2+(ω^((ω^2)∙2+ω∙2+1))∙2+(ω^((ω^2)∙2+ω∙2))∙3,3)
= H((ω^((ω^2)∙2+ω∙2+2))∙2+(ω^((ω^2)∙2+ω∙2+1))∙2+(ω^((ω^2)∙2+ω∙2))∙2+(ω^((ω^2)∙2+ω+3)),3)
= ...
Just post your solution.
Yeah, it gets horribly messy
I mean, its way bigger than Graham's number.
And Graham's number cannot be written in the observable universe.
Anywho, I'll be back later.
Feel free to ponder how large these numbers are and what sorts of nightmares must await if my TREE(3) program involves crazy symbols.
And welcome to Googology :D
(Also, H(ω^x,n) = f_x(n) in the fast growing hierarchy)
15:10
Write a program to calculate those things seems easy enough, and yet the result is extremely large. Nice.
You can also now understand how large some functions are.
 
3 hours later…
18:15
Hey @Mr.Xcoder
Hello @SimplyBeautifulArt
U gud with the above?
I am reading the above rn.
@SimplyBeautifulArt So $ε_0=\{ω\uparrow\uparrow 0, ω\uparrow\uparrow 1, ω\uparrow\uparrow 2, ω\uparrow\uparrow 3, ...\}$?
18:19
@Mr.Xcoder yup
@SimplyBeautifulArt Ok, I think I got the above, but could you please give me another exercise with the Hardy hierarchy? I want to test my abilities first (also typing dot product is annoying on my keyboard, so do you mind if I use another symbol for that instead?)
@Mr.Xcoder sure, I don't mind something like a period or something or implicit multiplication.
Try out H(ε_1,2)
Ok that's going to take a while
Just stop when you feel comfortable.
ಠ_ಠ $ε_1 = \{1, ε_0, ε_0^{ε_0}, ε_0^{ε_0^{ε_0}}, ...\}$
These numbers are just out of control
:P
18:25
that's the point :P
The goal: Write out everything that's sane, then go beyond. :D
Ok check the first part to make sure I am not dumb:
H(ε_1,2)=
H(ε_1[2],2)=
H(ω↑↑2,2)
Oh. Where did I go wrong?
3 hours ago, by Simply Beautiful Art
While we're at it, we may as well make
ε_1 = {1, ε_0, ε_0^ε_0, ε_0^ε_0^ε_0, ...}
ε_2 = {1, ε_1, ε_1^ε_1, ε_1^ε_1^ε_1, ...}
etc.
18:27
3 mins ago, by Mr. Xcoder
ಠ_ಠ $ε_1 = \{1, ε_0, ε_0^{ε_0}, ε_0^{ε_0^{ε_0}}, ...\}$
Yup I thought it was $0$, I am dumb
H(ε_1,2)=
H(ε_1[2],2)=
H(ε_0^ε_0,2)=
H(ε_0^(ε_0[2]),2)=
H(ε_0^(ω↑↑2),2)
@SimplyBeautifulArt Please check again, I am quite dumb as you can see.
Which is actually $H(ε_0^{(ω^ω)},2)$
Right?
@Mr.Xcoder Nah, its good
@Mr.Xcoder Yup
4 hours ago, by Simply Beautiful Art
Now we can get into some syntax.
Posting for easy access ^
@SimplyBeautifulArt I arrived to $H(ε_0^{ω\cdot ω},2)$. What should I do now?
Nvm...
@SimplyBeautifulArt Is the result a (relatively) small integer?
Also I need another check (especially the last steps), and I think I overcomplicated this:
H(ε_1,2)=
H(ε_1[2],2)=
H(ε_0^ε_0,2)=
H(ε_0^(ε_0[2]),2)=
H(ε_0^(ω^ω),2)=
H(ε_0^(ω^ω[2]),2)=
H(ε_0^(ω^2),2)=
H(ε_0^(ω*ω),2)=
H(ε_0^(ω*ω[2]),2)=
H(ε_0^(ω+ω),2)=
H(ε_0^(ω+2),2)=
H(ε_0^4,2)=
@Mr.Xcoder Definitely not
@Mr.Xcoder Almost.
H(ε_0^(ω+2),2)
ε_0^(ω+2) = a^(b+1)
So this was right: H(ε_0^(ω+2),2)?
Yup
You need to pull out the exponent into multiplication.
Wait can I just transfer the +2 to the ,2 part?
NVM ^...
@SimplyBeautifulArt Huh?
@Mr.Xcoder a^(b+1) = (a^b)a
Yeah but... my brain is slow
lol
ε_0^(ω+2) = (ε_0^(ω+1))ε_0
Now we can apply (ab)[n] = a(b[n])
Hey @RobertFrost
Wait I can handle it.
18:44
We be learning 'bout ordinals and how to make stupidly big numbers with them.
H((ε_0^(ω+1))ε_0,2)=
H((ε_0^(ω+1))ε_0[2],2)=
H((ε_0^(ω+1))ω*ω,2)
@SimplyBeautifulArt ^?
And then that again...
@SimplyBeautifulArt I have this. Won't we just loop like this forever?
@Mr.Xcoder ε_0[2] = ω^ω
Sorry, I'm in class, so I may be slow to respond.
And no
It will end eventually
@Mr.Xcoder On each step, the ordinal gets smaller and smaller...
@SimplyBeautifulArt Where did you pull ε_0[2] from?
4 hours ago, by Simply Beautiful Art
ε_0 = {1, ω, ω^ω, ω^ω^ω, ω^ω^ω^ω, ...}
(Note that a^b^c = a^(b^c))
@SimplyBeautifulArt Yes but where does ε_0[2] appear in our situation?
@SimplyBeautifulArt No problem
18:55
@Mr.Xcoder ((ε_0^(ω+1))ε_0)[2] = (ε_0^(ω+1))(ε_0[2])
Also I assume we can change the order since multiplication is commutative.
Also I have to say you are a very cool professor
@Mr.Xcoder No
Definitely no
@Mr.Xcoder I'll also have to mention I'm not even 18 yet.
18:58
@Mr.Xcoder Imagine ω∙2 versus 2∙ω.
@SimplyBeautifulArt ε_0^(ω+1))(ε_0[2])=ε_0^(ω+1))(ω^ω). And in H(ε_0^(ω+1))(ω^ω),2)=H((ε_0^(ω+1))(ω^ω))[2],2)=H(ε_0^(ω+1))(ω^ω[2]),2)=H(ε_0^(ω‌​+1))(ω^2),2) right?
@SimplyBeautifulArt You're like at IMO level :D?
ω∙2 = ω+ω = {ω, ω+1, ω+2, ω+3, ...}
2∙ω = {0, 2, 4, 6, 8, ...}
They are entirely different.
I understand.
@Mr.Xcoder :P
@Mr.Xcoder Yup
@SimplyBeautifulArt I am not joking, I am asking.
19:01
@Mr.Xcoder I'm not actually entirely sure what IMO level is.
Do you want to/Did you compete in IMO (international mathematics olympiad)?
If you mean university level mathematics, sure, but this usually isn't the sort of thing you'd find in the average university.
@Mr.Xcoder No, I don't do olympiads. Those don't really teach you 'new mathematics', which is what I'd like.
I understand, just being curious.
I have:
H((ε_0^(ω+1))ε_0,2)=
H((ε_0^(ω+1))ε_0[2],2)=
H((ε_0^(ω+1))ω^ω,2)=
H((ε_0^(ω+1))ω^2,2)=
H((ε_0^(ω+1))ω^2,2)=
H((ε_0^(ω+1))ω*ω,2)=
H((ε_0^(ω+1))ω*2,2)=
19:04
Looks good
For now. I have to go eat and will be back in ~5'
(using 2 instead of 3 simplifies things so much)
@Mr.Xcoder cya
@SimplyBeautifulArt Anyway, I am not sure on the steps to follow, so you can leave me a hint while I'm gone.
Sure. ω^2 = ω∙ω.
ω∙2 = ω...?
(Think about what x∙2 is in general.)
19:17
i am back
@SimplyBeautifulArt Umm, not sure what you mean by that
Oh of course $ω∙2 = ω+ω$
H((ε_0^(ω+1))(ω+ω),2)=
H((ε_0^(ω+1))(ω+ω[2]),2)=
H((ε_0^(ω+1))(ω+2),2)=
Now what?
@SimplyBeautifulArt TBH I have no idea how to continue.
19:37
Well
(ε_0^(ω+1))(ω+2) = (ε_0^(ω+1))(ω+1) + ε_0^(ω+1)
Oh well
x'D
It sure is a nightmare
Well
Time to go bigger?
@SimplyBeautifulArt No it's a daydream.
@SimplyBeautifulArt Yes.
19:40
Well
ε_(x+1) = {1, ε_x, ε_x^ε_x, ε_x^ε_x^ε_x, ...}
ε_x = {ε_(x_0), ε_(x_1), ε_(x_2), ε_(x_3), ...}
Wait, I am havibg trouble with the internet connection. Ping you when I move back to my mac, currently on mobile
@Mr.Xcoder I'm about to go, and I won't be back for about 3.5 hours.
@SimplyBeautifulArt Ok then, see yiu tomorrow then, I’ll most definitely sleep
Thanks a lot for this wonderful lesson! O/
In syntax,
ε_0[n] = ω^^n
ε_(x+1)[n] = (ε_x)^^n
(ε_x)[n] = ε_(x[n])
19:43
No problem. Hopefully you can imagine things like ε_(ε_0)
This
I can't think 4D :D
This is still smaller than TREE(3).
;)
@user202729 Oh yeah, FYI, this is all less than TREE(3)
@SimplyBeautifulArt How do you know? Do you compare them using the fast growing hierarchy?
But if you stick to it, I'll explain all the way up to TREE(3) and quite a ways beyond.
I will definitely stick to it.
19:46
@Mr.Xcoder TREE has an ordinal of approximately ψ(Ω^((Ω^ω)ω))
So... yeah
Cya!
Ω_Ω Cya!
20:03
@SimplyBeautifulArt hi, i only dropped in briefly earlier
20:40
@SimplyBeautifulArt For when you come back, can $$H(ε_0^{ω+1}(ω+1)+ε_0^{ω+1},2)$$ be worked on further? If so, how? I tried splitting, all the regular rules but I didn't succeed coming up with a method that simplifies it.
 
2 hours later…
23:06
@Mr.Xcoder You need to use (a+b)[n] = a+(b[n]) and a^(b+1) = (a^b)•a.
23:59
What we've done so far:
H(0,n) = n
H(x+1,n) = H(x,n+1)
H(x,n) = H(x[n],n)
(a+b)[n] = a+(b[n])
(a∙b)[n] = a∙(b[n])
(a^b)[n] = a^(b[n])
a∙(b+1) = a∙b+a
a^(b+1) = (a^b)∙a
ω[n] = n
ε_0[n] = ω^^n
ε_(x+1)[n] = (ε_x)^^n
(ε_x)[n] = ε_(x[n])
2

« first day (244 days earlier)      last day (155 days later) »