Tag Archives: information science

Unbaby Almost

About a month ago I was tickled to discover Unbaby.me, a plugin for Facebook that attempts to detect pictures of babies and replace them with an image feed of your choosing. It’s a neat idea, but the system fails on two accounts:

1) It traces words and phrases like “OMG SO CUTE!!! 🙂 🙂 :)” to determine if a picture is likely a baby. This takes out small furry animals as collateral damage, which I’m actually okay with, and also sometimes pictures of people. Sorry Ashley Bradarich, you’re still a celebrity but I love you for it. Your devil-angel photo is certainly TOO CUTE FOR WORDS HOLY GAWDS!!1!.

2) I set mine to relay the amazing website designs image feed on DeviantArt but since people there don’t understand what categories are I get results like the following:

See, this is my kind of information science. IR can be so much more fun than convoluted math equations.

Strange social norms

This post was a while back, but I juts now have gotten to it. A fellow PhD student, in an attempt to help us brainstorm more creative ways to recruit new information science students, sent us the following advertisement for reference:

As he explained, it’s a puzzle meant to draw you in:

In this poster, everything is hidden in the binary code (please ignore the girl at the background).

The decipher process can be found at http://xrl.us/bnx5jc (in Chinese). Here are the steps to get the material in natural language (I skip some trial and reasoning steps):

  1. OCR the binary code from the image.
  2. Save it as a binary file and name it bin.gz (you can get the name of the file by interpreting the binary code).
  3. Unzip the file and get a file called bin. This is a java class (again, tell from the binary code).
  4. Save the file as a java class and run it. You can get the correct name of the class (i.class) from the java error message.
  5. Run the class again and finally you will get an url: www.i.u-tokyo.ac.jp/fun/hikari-loveletter

What puzzles me about it is not actually the number cipher, which I imagine is pretty neato, but the the relevance of the woman in the background. Is she supposed to be just a pretty background, like a flower pattern? Are they assuming only heterosexual men apply and that this woman will get their attention? Is this like insurance companies that put naked ladies on billboards, but a subdued version? I’m not yet ready to label it sexist, simply because I don’t understand Japanese culture enough to comprehend the context. Right now I can’t help but see it as comically non-sequitur!

Common First Names

Many people have noticed I sometimes give people (or characters) strange (occasionally funny) nicknames. Well here’s to some justification, straight from Wolfram Alpha’s analysis of several thousand Facebook friends – common names:

Laura/Lauren 26
Sarah 19
Liz/Elizabeth 17
Katie/Katherine 16
Emily 13

Chris/Christopher 19
Mike/Michael 19
John 16
Matt/Matthew 16
Ben/Benjamin 14
Dan/Daniel 14
Dave/David 13
Brian 13

So, chances are if you have one of these names I’m probably going to have to make something more unique up for you. Sorry!