Many computer-based authentication schemata are based on pass- words. Logging on a computer, reading email, accessing content on a web server are all examples of applications where the identification of the user is usually accomplished matching the data provided by the user with data known by the application. Such a widespread approach relies on some assumptions, whose satisfaction is of foremost importance to guarantee the robustness of the solution. Some of these assumptions, like having a "secure" chan- nel to transmit data, or having sound algorithms to check the correct- ness of the data, are not addressed by this paper. We will focus on two simple issues: the problem of using adequate passwords and the problem of managing passwords. The proposed solution, the pathword, is a method that guarantees: 1 that the passwords generated with the help of a pathword are adequate (i.e. that they are not easy to guess), 2 that managing pathwords is more user friendly than managing passwords and that pathwords are less amenable to problems typical of passwords.