We expected Week 8 to be a productive-yet-uneventful, but, because every week holds surprises (that truth is indeed the dominant constant of ETC life!), the fact that some shifts in the direction of our project unexpectedly manifested was not altogether startling. Regardless, we pressed on within our mini-teams and continued to progress toward our persistent goal of establishing a fixed and efficient system of data collection and processing.
This week’s most significant news was that, on Tuesday, Ben informed us that the role of the custom Dead Space 3 level we’ve been planning on using as the basis of our user testing once Blade implemented our designs will likely be significantly diminished. Instead of using it as a focus of our testing, Ben suggests that we should test the majority of our users with existing Dead Space 3 levels, then have some testers in later sessions play our level, additionally. The rationale behind this thought is that Visceral Games, Dead Space 3’s developer, will be able to use our findings more practically if they stem from the release version of their product rather than a level specific and limited to our experimental effort. Our level might still prove useful to them, however, as we’ll now have the opportunity to share with members of their studio how testers respond to the segments of it designed to stimulate specific player type. Perhaps said information will prove enlightening, and perhaps not, but Ben’s other divulgence was that we’ll now be presenting the fruits of our research to Visceral Games at the end of the semester. We’re all thrilled at the opportunity, and sincerely hope that our work can positively influence future development there in even some modest way! The team’s still discussing and processing the implications of this recent information on our Halves presentation and the project entire, but we’re confident that Ben will assist and guide us to the best of his ability. Concerning more planned-for tasks, our Level and User Testing team executed their first official round of tests this week, and utilized a mixture of players from the public, fellow CMU ETC students, and EA employees. The results, as always, were compelling and revealing, and Martin and Nathan continue to refine the process of data-gathering. On the Data Team, meanwhile, Star sustained his exploration into correlation algorithms, and, with Emmanuel’s help, installed Mahout, Hadoop, and Unix. Shaveen and Vera further collaborated on database construction and refinement, while, respectively, Shaveen sought to better understand how to generate and gather Xbox 360 telemetry data and Vera began editing and processing the behavioral data videos gathered during user testing. Finally, Anabelle and Emmanuel, the Origin Team, presented their lovely and thoughtful recommendation system concepts to the rest of the full team, and to Ben. Individually, Anabelle’s been ensuring the graphical fidelity of our mock-up and its cohesion with Origin’s branding standards, while Emmanuel’s been building the functional HTML core atop which the graphics will be grafted.
Though next week is technically Spring Break, we plan to conduct Round 2 of our user tests and make as much headway as possible on both Data and Origin fronts. The week after next is (already!) Halves, and we plan to incrementally build and polish our presentation so that we can begin rehearsals.