|
normal contact posted:Even if pi contained every finite sequence of digits (which isn’t proven yet), there’s no guarantee that these two numbers would take up less space than any other compression method. I hope somebody quotes me and blows my mind with some current compression technology that can do something similar. edit; I am very wrong. Khorne fucked around with this message at 22:46 on Mar 22, 2018 |
# ? Mar 22, 2018 21:31 |
|
|
# ? May 29, 2024 23:10 |
|
Khorne posted:You only need the start position and a length, and there are tons of clever ways to store the start position in very few bytes. Citation needed.
|
# ? Mar 22, 2018 21:57 |
|
Salt Fish posted:Citation needed. There are endless ways to do this. You could even use the pi-compression algorithm iteratively if you wanna be cheeky about it, but you could also just use normal compression that we use today.
|
# ? Mar 22, 2018 22:08 |
|
Khorne posted:It likely would if the sequence were long enough. You only need the start position and a length, and there are tons of clever ways to store the start position in very few bytes. I'm not up to date on compression methods, but if you literally don't even have to store any data at all which would be like your compressed file passing a small 4-12 byte hash to a web api to get the uncompressed data. The question is how efficiently your computer can get to that sequence in pi from the calculated index, but that's not space related. It's literally impossible to compress data when every encodeable value is equally likely.
|
# ? Mar 22, 2018 22:08 |
|
XK posted:It's literally impossible to compress data when every encodeable value is equally likely. Technically, if pi behaved that way, there would be "correct" start positions that are larger than the data being compressed. Although I am unsure how the math works out for first occurrences, but for nth occurrences it would have to be the case. Khorne fucked around with this message at 22:17 on Mar 22, 2018 |
# ? Mar 22, 2018 22:10 |
|
Khorne posted:I get that both finding the right position for the sequence and then having that sequence available is not practical. I'm not saying it's impractical, I'm saying it's impossible. Thought exercise: Let's limit ourselves to encoding a start point from 0-7. That requires 3 bits. How do you compress that to less than 3 bits? Go to 2 bits, how do you represent 8 possible different pieces of data when you only have 4 different encodings? You can't without your compressed offsets becoming indistinguishable from each other.
|
# ? Mar 22, 2018 22:15 |
|
just decrease the font size, duh
|
# ? Mar 22, 2018 22:16 |
|
XK posted:I'm not saying it's impractical, I'm saying it's impossible. And the ways branch out from those ideas. But obviously, if your start position is larger than the data then it's all pointless. edit: I guess both methods would work in some cases but definitely wouldn't work in a general sense. Khorne fucked around with this message at 22:45 on Mar 22, 2018 |
# ? Mar 22, 2018 22:22 |
|
Don't you guys have contactless in the colonies? Also lol the bitcoin price graph is gonna end up looking like a loving seesaw isn't it.
|
# ? Mar 22, 2018 22:23 |
|
Khorne posted:Very few is kinda relative here. You need to be able to accurately represent two positive integers. The start position and the length. The integer is larger than the original content. Try to even think of a trivial case where you're not encoding the literal value of pi. How about the letter A for example, do this process for A.
|
# ? Mar 22, 2018 22:23 |
|
Salt Fish posted:The integer is larger than the original content. Try to even think of a trivial case where you're not encoding the literal value of pi. How about the letter A for example, do this process for A.
|
# ? Mar 22, 2018 22:24 |
|
Khorne posted:Very few is kinda relative here. You need to be able to accurately represent two positive integers. The start position and the length. mathematicians hate this secret trick!
|
# ? Mar 22, 2018 22:25 |
|
Khorne posted:Oh. I thought you meant the pi stuff. Yes, I was talking about encoding the starting offset. The issues of "sequence inside pi" thing are even worse. It's literally impossible to compress the starting offset. Provably impossible.
|
# ? Mar 22, 2018 22:27 |
|
Salt Fish posted:The integer is larger than the original content. Try to even think of a trivial case where you're not encoding the literal value of pi. How about the letter A for example, do this process for A. You can represent 1592 as 3,4. This kind of ignores the issue of the format needing to denote lengths or have a delimiter of some kind, too. It won't even be the same size for trivial examples. You could represent the entirety pi as 0,0 if you wanted 0 to be an otherwise invalid length. Khorne fucked around with this message at 22:33 on Mar 22, 2018 |
# ? Mar 22, 2018 22:29 |
|
Wtf is this nerd poo poo
|
# ? Mar 22, 2018 22:32 |
In bitcoin news: City of Atlanta's computer systems are being held ransom for bitcoin. This is good for bitcoin?
|
|
# ? Mar 22, 2018 22:35 |
|
XK posted:Yes, I was talking about encoding the starting offset. The issues of "sequence inside pi" thing are even worse. Khorne fucked around with this message at 22:46 on Mar 22, 2018 |
# ? Mar 22, 2018 22:39 |
|
Khorne posted:It likely would if the sequence were long enough. You only need the start position and a length, and there are tons of clever ways to store the start position in very few bytes. I'm not up to date on compression methods, but if you literally don't even have to store any data at all which would be like your compressed file passing a small 4-1024 byte hash to a web api to get the uncompressed data. The question is how efficiently your computer can get to that sequence in pi from the calculated index, but that's not space related. You cannot. The field that proves this is called information theory. Essentially, you'd need at least 2^n starting positions to encode all sequences of length n. You need n bits to encode numbers up to 2^n. Some offsets or bitpatterns you could encode more efficiently at the cost of encoding other less efficiently. Regular compression essentially does this in a clever way (for simple examples, see run-length encoding or hamming coding) so what you want to represent has a short representation, but using positions of pi makes the encoding random, so it's a toss-up whether you win or lose. Which is still better than bitcoin, where you can only lose, so there's that.
|
# ? Mar 22, 2018 22:43 |
|
pi doesn't even have all the digits from 0-9 until 33 digits in, so it can't even represent single digit numbers efficiently. it gets worse from there.
|
# ? Mar 22, 2018 22:46 |
|
klafbang posted:Some offsets or bitpatterns you could encode more efficiently at the cost of encoding other less efficiently. Regular compression essentially does this in a clever way (for simple examples, see run-length encoding or hamming coding) so what you want to represent has a short representation, but using positions of pi makes the encoding random, so it's a toss-up whether you win or lose. After thinking about things and some quick reading I was very, very wrong. I've never really dealt with encoding directly. Especially not in a theoretical sense. Thanks for the posts guys, I learned something in the bitcoin thread. toiletbrush posted:pi doesn't even have all the digits from 0-9 until 33 digits in, so it can't even represent single digit numbers efficiently. it gets worse from there. edit: I like this post about why it sucks. But also, there are simple reasons like "there are plenty of algorithms better than the digits of pi". I've also skimmed an information theory textbook and learned quite a lot. Khorne fucked around with this message at 23:05 on Mar 22, 2018 |
# ? Mar 22, 2018 22:49 |
|
The pi compression topic make me nostalgic since it's one of the first things I thought of when getting deeper into qbasic in middle school, and I bet that like 90% of all computer touchers had the same idea when they were kids
|
# ? Mar 22, 2018 22:58 |
|
Sentient Data posted:The pi compression topic make me nostalgic since it's one of the first things I thought of when getting deeper into qbasic in middle school, and I bet that like 90% of all computer touchers had the same idea when they were kids Pretty sure I dismissed it on the grounds of calculating pi stuff to be impractical in itself. I never really thought about the realities of the offset. Khorne fucked around with this message at 23:10 on Mar 22, 2018 |
# ? Mar 22, 2018 23:06 |
|
Local Japanese media outlet Nikkei reported March 22 that Hong Kong-based Binance, which is currently the world’s largest exchange by trade volume, was to be ordered to shut down operations in Japan for failing to register with the Financial Services Authority (FSA). https://www.bloomberg.com/news/articles/2018-03-22/bitcoin-falls-after-report-that-binance-faces-warning-in-japan
|
# ? Mar 22, 2018 23:26 |
|
Khorne posted:This post made me realize I've thought of it before, I can even recall the computer I was sitting at in the 90s, and realized you can't do it. Yet somehow sitting here right now I argued you could in the hopes people would post mathematical proof you can't. Welp. Older & dumber. Yet you should be congratulated because you've put more thought and effort into the idea than most cryptocurrency creators put into theirs And when you found your idea wanting you discarded it instead of insisting that it would revolutionize everything
|
# ? Mar 22, 2018 23:29 |
|
Sentient Data posted:The pi compression topic make me nostalgic since it's one of the first things I thought of when getting deeper into qbasic in middle school, and I bet that like 90% of all computer touchers had the same idea when they were kids I once invented encryption by xor-ing with a one-time pad generated using a seeded PRNG. Luckily, Satoshi invented real cryptography shortly thereafter.
|
# ? Mar 22, 2018 23:34 |
|
Risc1911 posted:Local Japanese media outlet Nikkei reported March 22 that Hong Kong-based Binance, which is currently the world’s largest exchange by trade volume, was to be ordered to shut down operations in Japan for failing to register with the Financial Services Authority (FSA). this thread makes me laugh out loud in real actual life more than any other thread on Something Awful
|
# ? Mar 22, 2018 23:53 |
|
Burt Sexual posted:Wtf is this nerd poo poo bitcoin
|
# ? Mar 23, 2018 00:57 |
|
Risc1911 posted:Local Japanese media outlet Nikkei reported March 22 that Hong Kong-based Binance, which is currently the world’s largest exchange by trade volume, was to be ordered to shut down operations in Japan for failing to register with the Financial Services Authority (FSA). This is good for bitcoin.
|
# ? Mar 23, 2018 01:12 |
|
normal contact posted:Even if pi contained every finite sequence of digits (which isn’t proven yet), there’s no guarantee that these two numbers would take up less space than any other compression method. For any nontrivial text those two numbers almost certainly take more space than could be possibly held if every atom of the universe were converted to the most dense storage medium theoretically possible.
|
# ? Mar 23, 2018 01:21 |
|
Khorne posted:This post made me realize I've thought of it before, I can even recall the computer I was sitting at in the 90s, and realized you can't do it. Yet somehow sitting here right now I argued you could in the hopes people would post mathematical proof you can't. Welp. Older & dumber. Imagine.... meanwhile a man named Satoshi Nakamoto sat at a computer and gave humanity a new way to live and prosper through the blockchain. Makes u think
|
# ? Mar 23, 2018 01:30 |
|
https://twitter.com/maxwelltani/status/976926310445584386 Now that he is no longer relevant, Steve Bannon has to find a new bunch of weirdos to ingratiate himself with.
|
# ? Mar 23, 2018 01:38 |
|
Risc1911 posted:Local Japanese media outlet Nikkei reported March 22 that Hong Kong-based Binance, which is currently the world’s largest exchange by trade volume, was to be ordered to shut down operations in Japan for failing to register with the Financial Services Authority (FSA). Japan banning bit nonce!
|
# ? Mar 23, 2018 01:39 |
|
Bip Roberts posted:Japan banning bit nonce! it's racist of japan to ban beyonce just because she's chinese imo
|
# ? Mar 23, 2018 01:55 |
|
I would blow Dane Cook posted:https://twitter.com/maxwelltani/status/976926310445584386 This dude just seems to live tweet other news outlets stories and has acne.
|
# ? Mar 23, 2018 02:58 |
|
Khorne posted:This is actually true and simple to figure out. Oops. I could really blow your mind with my "All positive numbers added together equal -(1/12)" post.
|
# ? Mar 23, 2018 02:59 |
https://twitter.com/cz_binance/status/976997061156519937 so.. confirmed?
|
|
# ? Mar 23, 2018 03:32 |
|
So yea, pedocoins.
|
# ? Mar 23, 2018 03:39 |
|
I am beginning to think that instead of the massive, hour-long crash we have all been waiting for the hodlers and fodlers will just be cashing out over the months and years as they grow weary of the realisation they will never become the titans of industry they so desire for leaving their computers on all night. *sigh*
|
# ? Mar 23, 2018 04:22 |
|
klafbang posted:You cannot. The field that proves this is called information theory. What you need to do is jerk of two dicks since you have two hands... and what if you lined up the dicks so the up-stroke on one dick is the down-stroke on the other dick! FOUR DICKS!
|
# ? Mar 23, 2018 04:27 |
|
|
# ? May 29, 2024 23:10 |
|
Jombo posted:What you need to do is jerk of two dicks since you have two hands... and what if you lined up the dicks so the up-stroke on one dick is the down-stroke on the other dick! FOUR DICKS! https://i.imgur.com/Un2DKD1.mp4
|
# ? Mar 23, 2018 04:37 |