Abstract
We introduce in the present paper an operationally defined subclass within the genre of detective stories, specified on the basis of the logic programming model adopted in our Logtell project. Special attention is given to the treatment of communicative events. An SWI-Prolog plan-based tool was developed to compose consistent plots, conforming to the conventions of the genre. Seven criminal cases generated by the tool are described as illustration. It is shown, using the PlotBoard interface, how to run plan-based composition in interactive stepwise mode. We have also developed a storytelling system capable of representing the stories in the format of interactive comic books on tablet computers.
Highlights
The basic objective of this paper is to propose a strategy to define a subclass of the genre of detective stories, on the basis of a logic-programming model introduced in [1] and more rigorously formulated in [2]
3. the factors driving the characters to act. Speaking, such threefold specifications determine what kind of facts can hold in each state, what events can change a state, and what goals can motivate the acting characters to trigger events that compose consistent detective story plots
Choosing different initial states is a way to achieve a varied repertoire within a given genre – which is further enlarged by both (i) the ability of the plan-generator to backtrack to produce different plans to reach the same goals and (ii) the user-interaction features that can be built into the algorithm
Summary
The basic objective of this paper is to propose a strategy to define a subclass of the genre of detective stories, on the basis of a logic-programming model introduced in [1] and more rigorously formulated in [2]. One may object to the use of plan-generation to help composing stories, even with the user's interaction, on the grounds that the resulting plots would be essentially predictable. To this we would reply that, while never pretending to equal the creativity of talented writers, planners – given their ability to systematically explore the consequences of rule-based specifications – can often come up with the unexpected, as will be exemplified at the end of part A of Section II. A fuller description of our logic-programming model, which adopts the Entity-Relationship model [6] to specify facts, the STRIPS method [7] to define events, and situation-objective rules to motivate the characters' behaviour, as well as details about the system's implementation, can be found in a previous technical report.
Talk to us
Join us for a 30 min session where you can share your feedback and ask us any queries you have