This topic contains 112 replies, has 2 voices, and was last updated by GreenPixel 3 weeks, 5 days ago.
October 17, 2018 at 4:41 am #30852
I think the confusion here comes from the term “search”. It’s not actually “Searching” for your text. All this text isn’t saved somewhere before you start. It’s just an algorithm that can convert a fixed-length block of text into an index, and vice versa. The block of text required for the algorithm to operate is a specific length, so to run the algorithm and generate the index, it first has to fill in the remaining space with junk: spaces, random characters, or words. It gives you all three variants.
So when you run a “search”, it takes your string, the site randomly generates some buffer junk around it on both sides until the block of text is the right size, then the algorithm converts that to your index.
This is the reason why:
1. The text is always in the middle of the page.
2. The words you entered are surrounded by garbage characters in the “characters” result variant, even though your text seems out of place there.
3. If you were to search the same string twice, or search the string that someone else sends you, you’ll get a different index. The site is randomly generating that buffer, so when “searching” the buffer text that it creates will be different every time.
This doesn’t mean it’s fake. Had you randomly fallen upon that same index before you ever searched it, it would have lead to the same text, because the algorithm is deterministic in how it converts indexes to text blocks and back.October 17, 2018 at 4:20 pm #30870
Yes, that’s exactly how the search feature works. It’s been explained a few times, even by me, but unfortunately it gets buried in these forums. I only wish there was a more extensive write-up on one of the info pages, but the VSauce video covered it well (which is where I assume most of the traffic here comes from).October 17, 2018 at 4:27 pm #30871
So it’s not searching anything… it’s a generator! Yet again Basile’s fraudulences are exposed. To search a database of that magnitude would require nuclear power!October 17, 2018 at 5:25 pm #30875
It has literally never been a secret that the pages are generated. Jonathan has stated multiple times that the pages are generated. Here are some replies:
It’s also mentioned in Grains of Sand (4th paragraph), albeit not as technical. Not sure what your point is.October 17, 2018 at 6:25 pm #30876
My point is, it is called the “search” button! Cold lies!October 26, 2018 at 3:41 pm #31127
sir/mam when i am entering something in search option a messed up english letters are coming how could i know what is the meaning of that para or sentence i am just finding my particular sentence nothing else.
so, i am not able to know anything except that my typed sentence.
plz help, so that i can efficiently use thisOctober 26, 2018 at 4:38 pm #31129
There are other options if you look further down. The first one is “exact match” which just finds your search with the rest of the page consisting of spaces. The next one is “with random characters”, which fills the rest of the page with randomness. The one after that is “with random English words”, which fills the rest of the page with words. The last one is “title match”, which searches for books titles that match your search term (can only be seen in the first page, at the top in bold).November 13, 2018 at 10:18 pm #31536
For your second point, it’s just statistics. There are millions of pages with garbled nonsense text, with occasional pages with readable text scattered throughout. The algorithm can of course find a match in the library, because that’s what the library is for, but your matching text is nestled within millions of gibberish pages. If you were clicking through pages, it is extremely unlikely you would find anything readable at all.