User:Danny van der Kleij/how why

Free wikileaks pirate radio

How:

To successfully broadcast Wikileaks content at the Speed show, I had two requirements. One the process had to be automated as I didn’t want to present a performance piece. Two the project had to be run on the Belhuis computers where the Speed show was held. I opted for a Firefox plug-in that would automate the process, which I wrote using Greasemonkey, Jquery and Javascript. The script would redirect the Wikileaks content to the Google translate text-to-speech website, so that it was spoken out automatically once you visited the Wikileaks page. For transmission I used a standard FM transmitter, normally used in cars to broadcast from your mp3 player to your car radio. This transmitter was simply connected to the speaker port of the computer.

Why:

At the time there was a lot of discussion and on the taking down or backing up of the Wikileaks servers. Which I thought was rather trivial since the wikileaks content had been up for a while. In my opinion it was the same as taking down the original source of a pirated movie. Once the movie has been downloaded by someone, it can be reproduced and distributed again. Because of the reproducibility of all digital material, the source in essence loses its value. If the Wikileaks server would ever be taken down, the holders of downloaded Wikileaks material could recreate it. The concept of medium was also very important to me. Nowadays the average computer user also has access to some form of printer, thus if anyone had downloaded Wikileaks content it could be printed. I chose radio because it is one of the most fleeting media, it happens only at a certain time but it can be picked up by anyone within transmission radius.

More of the same

How: In this project I made use of two databases, namely the music recommendation website last.fm and the video website youtube.com. Once you enter the More of the same website you can fill in the music that you like. The python script that powers this website first checks if that search query is in the last.fm database. If it is in the database it searches for similar artists, and picks the least similar one. From this artist the script searches for keywords the artist has been tagged with. It takes one of these keywords and searches for a new list of artists associated with these keywords according to the last.fm database, again picking the artist that is the most unrelated. From this artist the script looks for a last.fm user that has been listening to this artist. The script then searches this users music listening history and picks the least listened to artist. This artist is then searched on Youtube.com. The more of the same website then displays all the complete llists it has searched for in a line of thumbnail size pictures of artists usually between 100 and 200 results, accompanied by the Youtube video.

Why: I wanted to make a statement against music recommendation websites. Music recommendation websites are made to recommend their users music that they like but the way they are constructed means that everything is based on keywords, genres and other peoples listening behaviour. This results in getting a lot of the same music, mostly music that you already know. More of the same does almost that, it presents the user with an overload of music recommendations, mostly based on genre and keyword similarities. However there are moments where is it falls of the map completely, presenting you with music that is somehow an error of the last.fm database or music that is made by amateurs that somehow made it’s way unto last.fm or youtube.com. This has led to More of the same to be a very surprising and sometimes very odd user experience.