While this might not be as strict (or serious?!) an evaluation technique as, say, heuristic analysis or usability testing, having two people act out the roles of the user and the user interface (UI) of system can, I think, be very revealing.
It’s a little bit like some of the more elaborate “paper prototyping” scenarios I’ve seen, but I first heard of this in a talk by Stephen P Anderson, at UXLx 2010. Perhaps the best thing to do is see Stephen describe it at another event in Norway.
A quick overview
It’s quite simple really… One person plays the role of the UI of a system, and their spoken responses are essentially scripted according to the text or options provided in the interface. The other person plays the role of a user with a given task. The two then interact, and the audience (developers, stakeholders, etc) listen to how that dialogue sounds and where the user begins to perhaps struggle with the interface.
This can then prompt discussion of how we might “humanise” the interface a bit more, to make this dialogue flow better, and thus to provide a better user experience overall. I like this because it reminds me of the way that Caroline Jarrett and Gerry Gaffney talk about form design in terms of a dialogue. The interface needs to “ask” question of give options in a way and using language that makes sense to the person using it. I think this is true of just about any interface, and serves as a good paradigm for evaluation.