Evren wrote: "Rather the community that the user is a member of, the rules of that community and the roles of that community should also be considered for HCI design and evaluation. (I think Activity Theory is quite well known in CSCL field and everybody knows the triangle developed by Engestrom, 1995). Following this shift in HCI research it is also important to reconsider usability testings."

I agree with Evren that activity theory can be used (and actually is already used) as the background theory in usability testing. However, in think is very difficult to decide how much stress do you put to the "mediating artifact" (= the computer / software) in the usability testing relying on the activity theory. How much the tool (the mediating artefact) should take care of the relations between the other elements of the activity system (subject, object, rules, community and division of labour)? I think it is important to realize the activity theory talks about whole system. Based on this you may of course analyse how well the "mediating artifacts" (the tools, the software) fits to the system. The problem comes that all the systems are unique and the elements are dependent on each other and are in interaction.

With the Fle3 CSCL software development we have try to design an artefact (=software) that works in a system where the other elements of the activity system are certain, too. In a way there must be a certain subjects, objects, rules, community and division of labour in place and only after this the Fle3 software is useful = the usability is high. And this is trues the other way round, too: if the elements are not there the software is not useful (the usability is low).

Simplifying you could say that you need a right tool to the task you are doing. Hammer is better for nailing than a screwdriver.

I also think that building an efficient activity system you must start from one of the elements. I also believe that the mediating artefact can be a good first "agent" to put the other pieces in a right place, too.

Of course in usability there is a lot of simple usability engineering things which must be met, as well. In way this part is naturally much more simple than when talking about the software as a part of some activity system.