The Watch-and-Comment paradigm has been proposed as the seamless capture of comments made by users when appreciating a video, which is watched using a computer or a TV set, for example. The comments can be associated with the original media in several ways so as to generate interactive videos automatically -the interactive video corresponding to the original video annotated with the captured comments. In this paper we explore the application of the Watch-and-Comment approach in situations in which the annotations are represented by operators used for the discrimination of instants and time intervals within continuous media. This particular type of annotation can be performed collectively by several users, over distinct instances of media and, for recorded media, at different times. Once these discriminations are available, combinations of the annotations may be employed by several classes of applications such as those which exploit collaborative selection, voting, cutting, classification and evaluation. We show the feasibility of the operators by implementing a proof-of-concept prototype on a Digital Television platform: the prototype supports multiple geographically distributed users who, using distinct input devices, perform collaborative tasks not planned in the original interactive-TV applications.