False friends constitute a source of error for learners of a language and for translators. However, false friends are constantly created either as loan words or as words which share a common derivation. This article investigates the phenomenon of false friends from a positive perspective in order to establish a classification towards a descriptive study of computer-related translations of English into Spanish. The aim is to show to what extent false friends is a widespread phenomenon when a word is borrowed from another language, or when two culturally and historically-related languages such as English and Spanish interact. The main task is to make the most of the semantic and pragmatic information provided by false friends and analyse actual occurrences of false friends in computer magazines as well as in computer science textbooks. As a result, different categories of false friends could be established.