Concurrent constraint programming (CCP) has emerged as a simple but powerful paradigm for concurrent systems; i.e. systems of multiple agents that interact with each other as for example in a collection of music processes (musicians) performing a particular piece. The ntcc calculus is a CCP formalism for modeling temporal reactive systems. In ntcc, processes can be constrained by temporal requirements such as delays, time-outs and pre-emptions. Thus, the calculus integrates two dimensions of computation: a horizontal dimension dealing with partial information (e.g., note > 60) and a vertical one in which temporal requirements come into play (e.g., a process must be executed at any time within the next ten time units). We shall show that the above integration is remarkably useful for modeling complex musical processes, in particular for music improvisation. For example, for the vertical dimension one can specify that a given process can nondeterministically choose any note satisfying a given constraint. For the horizontal dimension one can specify that the process can nondeterministically choose the time to play the note subject to a given time upper bound. This nondeterministic view is particularly suitable for processes representing a musician's choices when improvising. Similarly, the horizontal dimension may supply partial information on a rhythmic pattern that leaves room for variation while keeping a basic control. We shall also illustrate how implementing a weaker ntcc model of a musical process may greatly simplify the formal verification of its properties. We argue that this modeling strategy provides a "runnable specification" for music problems that eases the task of formally reasoning about them.