r.sine is a website that displays a random image or gif each time the page is refreshed. The site can be accessed at https://r.sine.com/index. However, while poking around, I realized that if you replace "index" with other words, the images are no longer random. We have 3 possible cases. Either the word displays random images (e.g. https://r.sine.com/index ), or the word displays a single image (e.g. https://r.sine.com/admin), or the word displays a "buffer" image which is used for several words, probably when the word is not addressed to an image. I decided to use a script to discover all the possible words to use. Yes, it's stupid. Welcome to the internet.
So it's a pretty straightforward process. We know that "index" offers random images (Cond1), that some words link to the same photo (Cond2) and that some words are only affiliated with a single image (Cond3).
- Cond1: random image →
random_pictures.txt - Cond2: buffer image used for several words →
same_pictures.txt - Cond3: single image for one word →
unique_pictures.txt| single image for 2 or more words →unique_pictures_2words.txt
So basically, we make requests for each word in the dictionary !
For this, I use either Wikipedia's word frequency lists
or words.txt from english-words GitHub Repo.
The word list I'm using has a problem... Those are only lowercase words. And certain words in lowercase do not give the same image as in uppercase so my script obviously tests both versions.
| https://r.sine.com/science | https://r.sine.com/SCIENCE |
|---|---|
Then I make a first request for a specific word and retrieve the image source code. I make a second request with the same word. If the 2 source codes are not similar, then we're in the case of the word "index" for example, where we'll only have random images (Cond1). If the 2 source codes are similar, this means that the image is necessarily identical and that the word leads to a specific image. Then, if the 2 source codes are identical, I check whether the latter is similar to a source code already known for the other words already screened. If it's identical, then the word leads to the "buffer" image (Cond2). If not, the word leads to a unique image (Cond3).
After that, I made the script to limit the number of requests, but that's a technicality that no one cares about.
Just replace index from https://r.sine.com/index with an other word (e.g. advice) → https://r.sine.com/advice
For V4, I use EazyPeazyE suggestion by using https://r.sine.com/index?html=true. With this technique, below the photo we have the name associated with the image. I have with a threading of 8, I reloaded the page for 24 hours and retrieved all the words that came out and the source codes associated with the images to sort them as explained later.
output (and version) are the post-processing folders. Sometimes my script makes errors because it goes too fast for server responses I think. So I go through the word list several times to be sure and I have a False Discovery Rate (FDR) folder to test it.
Here, I explain in more detail my reasoning and how I sorted.
-
Cond1: random image →
random_pictures.txt(e.g. https://r.sine.com/index for legal reasons, I don't display random images) -
Cond2: buffer image used for several words →
same_pictures.txt(e.g. https://r.sine.com/nicetry) -
Cond3:
-
single image for one word →
unique_pictures.txt(e.g. https://r.sine.com/admin) -
single image for 2 or more words →
unique_pictures_2words.txt(e.g. https://r.sine.com/bird and https://r.sine.com/parrot)https://r.sine.com/bird https://r.sine.com/parrot
-
Statistics of V1:
- 370101 words tested
- 1 word does random pictures → index
- 1531 words lead to a unique things:
- 1455 words do a unique images
- 29 images accessible via 2 words (58 words)
- 2 images accessible via 3 words (6 words)
- 11 words are specials... web page, redirection to youtube, snippet of r.sine code, pdf, broken images (repaired in html), mysteries
Statistics of V2 📈 (2.7x more words analyzed and 185 new words found):
- 1000316 words tested:
- 1 word does random pictures → index
- 1716 words lead to a unique things:
- 1631 words do a unique images
- 32 images accessible via 2 words (64 words)
- 2 images accessible via 3 words (6 words)
- 15 words are specials... web page, redirection to youtube, snippet of r.sine code, pdf, broken images (repaired in html), mysteries
Statistics of V3 📈 (2x more words analyzed and 24 new words found):
- 1000316 words + 1000001 number (0-1000000) tested:
- 1 word does random pictures → index
- 1757 words lead to a unique things:
- 1671 words do a unique images
- 32 images accessible via 2 words (64 words)
- 2 images accessible via 3 words (6 words)
- 15 words are specials... web page, redirection to youtube, snippet of r.sine code, pdf, broken images (repaired in html), mysteries
Statistics of V4 📈 (6094 new words found):
Thanks to EazyPeazyE by using https://r.sine.com/index?html=true, we have access to the name of the picture.
- 1000316 words + 1000001 number (0-1000000) tested:
- 1 word does random pictures → index
- 7851 words lead to a unique things:
- 7033 words do a unique images
- 335 images accessible via 2 words (670 words)
- 29 images accessible via 3 words (87 words)
- 9 images accessible via 4 words (36 words)
- 25 words are specials... web page, redirection to youtube, snippet of r.sine code, pdf, broken images (repaired in html), mysteries
For those who are curious, the picture buffer corresponds to the words calm and noneed
Note 1: some words are internet pages. They are no longer executable on r.sine.com but I host them (on Netlify) so that you have access to them.
Note 2: since certain words do not lead to images, in the special_words folder you will find the source code given by the word in question (in html or txt or pdf version)
A scientific chicken article in PDF format. Thank you r.sine.com 😅
Music group promotion web pages ?
It's a comic strip on a white background. The white background and the way the GIF is generated are quite atypical.
Web page to comment on photos of Rachel Sterne Twitter ?
Images that no longer work on r.sine.com. HTML work.
A video of Peppa Pig 🐖
A mystery... I have no idea about it
Except for robots: https://en.m.wikipedia.org/wiki/Robots.txt
I think this is a reference to the method to prevent Google bots from scanning the site. That makes rsine a deep web site, I guess
I don't know if it's a bug or if it's intentional. It's a piece of code that helps us understand how images are returned to us on https://r.sine.com/index.
🎸 Rathooooolllleeeee... 🎵
🏓
Leads to https://www.youtube.com/watch?v=yoy4_h7Pb3M (html broken but works on https://r.sine.com/topmen)
These words are broken and we have no source code for the image so probably the photos no longer exist
Access forbidden, Oddly, with index we can have access to them but why give the possibility of coming across these pages by chance?
Request more investigation. Oddly, with index we can have access to it but I don't have the source code but if we search for it manually I can't find them
I think we found everything, then if we looked carefully... the creator left a clue 😉 but you won't find it in this GitHub 😉
The images on r.sine.com may offend the sensibilities of users. If you do not have the majority (of your country), I disclaim all responsibility for what you find on this site. This GitHub project is just a curiosity about another Internet curiosity.