A CAPTCHA in the Text Domain

Research on CAPTCHA has led CAPTCHA design into adopting almost exclusively graphical implementations that deal mostly with character recognition. This has reached an exhaustion point, where new approaches are vital to the survival of the technique. This paper discusses the early stages of a research that intends to solve the open problem of a CAPTCHA in the text domain offering, this way, innovative research possibilities to the CAPTCHA paradigm. It is essentially an investigation on a CAPTCHA that draws its security from the cognitive and computational aspects behind phonetic punning riddles found on Knock-Knock Jokes. By the specification of a computational model, the implementation of a prototype and its experimentation with human individuals, it is shown that the proposal is indeed feasible and that studies in non conventional areas for Information Security are the key for developing the proposed goal.