Proceedings of the Annual Meeting of the Cognitive Science Society

Behavioral studies often warrant the inclusion of computa-tional participants in addition to humans. However, connect-ing computational cognitive or AI frameworks to GUI-basedsoftware developed for human use is extremely difficult. Thisresults in researchers either (1) diving into software code toappend an API for computational participants, (2) developingtwo separate versions of task code – one for human and one forcomputational participants, (3) cherry-picking research tasksthat already include both a GUI and an API, or (4) finding away to publish the research “as is” without the potentially use-ful results from running simulated participants on task. Theseemingly minor nuisance of the API-GUI dichotomy in to-day’s world of software development is, in fact, responsiblefor reduction in scientific progress. This work proposes afunctional-essence approach to software development, and theuse of STAP (Simple Task-Actor Protocol) as a standard UIinteraction language, for overcoming the API-GUI dichotomyand enabling access to the same software for both human andcomputational participants. We envision the adaptation of theproposed methodology to enable selection of off-the-shelf be-havioral tasks, decorative templates, and cognitive/AI frame-works for a more efficient path to research results.