Hard to believe
This topic contains 6 replies, has 1 voice, and was last updated by Ray 6 months, 2 weeks ago.
April 2, 2017 at 1:05 am #14508
Either I still don’t get the concept or it seems fake/useless.
I searched a long sentence, copied it and searched it several times again and again.
Every single time the result pointed me to different places. I understand that my sentence may appear in various contexts. But why does the same search generate different results?
Also, it seems to find even very long sequences of words when searched, but browsing through multiple existing pages almost never yields even a short sequence of 3 or 4 real words.
Can someone explain the first observation and did anybody succeed to find something new and meaningful just browsing?April 2, 2017 at 1:34 am #14510
The way search works I believe is that it takes your string (such as “Dave”), and then fills the space around it with random characters or words (so “asfasjfajfwuaDaveiaidfjwjiafji”).
From here, the site can trace the algorithm back to find the exact places that a specific string of random character surrounding your query would appear. The concept that drives this is “encryption and decryption.”
It’s important to note that, while every single page is possible, not every single book is possible with the current hexagon character cap. It’s also important to note that specific pages are not unique, and repeat several times within the current limits.April 2, 2017 at 12:25 pm #14524
My question is: Is an actual virtual library already there where all possible 3200 or so long texts are stored or does it create the text only when you search for it by put it in the search box? To me it would make a big difference. The former would really constitute a fascinating resource to work with, the latter would be just a simulation.
Or putting it in practical terms: would I find something meaningful there that no human has ever typed just by browsing?April 2, 2017 at 1:24 pm #14525
Listen up nigga, storing the entire content of this fucking library is basically impossible, probably the amount of the data printed on standard A4 papers won’t even fit into the entire known universe. There’s a DICKLOAD of data outta there, believe me.April 2, 2017 at 2:15 pm #14528
Watch your language, you with the stupid name selection.April 3, 2017 at 2:17 am #14544
It would take too much space to store. The pages are not, however, “generated” upon search as you seem to describe it.
If you browsed, the hex/wall/shelf/volume/page would all be put through the algorithm. It would then return to you the exact text that would appear on that page. This text will always be the same, everytime you go to that location.
So, the library is “simulated” in the sense that it is not physical, but it retains ALL qualities of the actual library. There is indeed an astronomically low chance that you can find a sentence by browsing.April 3, 2017 at 2:20 am #14545
And the randomly generated text put around your search query is NOT placed on the page. Instead, the algorithm is reverse engineered to find where your query + random text is in the library.
It is indeed a search, the random text that it generates is simply because the algorithm cannot be reverse engineered without already having a FULL PAGE of 3200 characters to look for. Without max characters, the algorithm cannot be reverse engineered.