I got back home. The main reason for return was the classes in school, I can't afford to miss them. During one of the classes, we had to do an assignment where we should analyze the influence of a parameter to a document classification performance. Everybody was doing it manually and took only few samples. And everybody was using Windows. But I had my laptop with Linux with me. Since the document classification program accepted command-line parameters, we (me and my colleague) wrote a small shell script which could take any number of samples, imported data to R and created a quite precise plot. When the plot was ready, the teacher looked at it and said: I've never seen such plot before. It was a compliment, I suppose.