Iterative design allowed us to integrate user feedback into our interface, and justify the changes we needed to make. If we did it again we would re-focus on making sure our users got a consistent experience from the testing facilitator; sometimes the roughness of our testing template negatively impacted the testing experience, it confused the test facilitator and he found himself needing to make changes to it on the fly.
If we did it again we'd go with a framework that encapsulated some of the MVC structure. Our code rapidly grew very large, and implementing the features we needed grew exponentially harder as we lost track of what code manipulated what visual element. We considered the framework too risky a change to make after the work we had already completed for GR3; none of us felt motivated to rewrite what had already been written, both due to time constraints and fear of breakage.
For our final user testing evaluations we attempted to remotely run a test - by employing FaceTime and Google Hangouts Screensharing. We felt that the fidelity of the observations we made suffered as a result. Screencasting software did not let us physically guide the user in case he or she got completely confused and lost. (We were relegated to attempting to describe the look of the thing to click on next.) It also proved difficult for two people (a facilitator and a recorder) to process the data received from the remote user - we recorded notes in a text document on the same screen that the screencast was playing on. Looking back, taking notes by hand would have made things slightly easier.
We evaluated the results of the user tests by examining the emotional state of the user (frustrated, happy, etc), how easily they were able to accomplish the goals we outlined, and by observing the way their workflow moved through our interface.