Summary: Comparative usability evaluation studies, like the CUE studies, have shown that independent evaluations of the same product produce only limited overlap in terms of the number of issues reported in common. This lack of overlap is due to the very human variety in the perspectives of users and evaluators. These findings do not, however, invalidate usability evaluation methods: individual teams of evaluators are finding real problems, just not all of them. Usability professionals should exploit the same variability between people that leads to the limited overlap: instead of comparing the results of individual evaluations, one should be pooling them together! A greater percentage of all usability issues can be identified if more frequent usability evaluations are used throughout product development and more varied perspectives, namely more usability professionals, more tasks, and more (or simply more heterogeneous) participants, are included in evaluation.

Summary: CAPTCHA tests require users to identify characters in a sequence that has been distorted. While these distortions are necessary to throw off optical character recognition programs, they also increase the likelihood of human errors. Knowledge of how letter and word recognition happens in the brain can help decide which CAPTCHA tests will be most manageable for users. In particular, associations in the brain between representations for letters and words make humans more accurate at recognizing letters in the context of words and pseudowords than random strings. In addition, since there seem to be separate brain mechanisms that handle digits and letters, it is not advisable to use both kinds of characters in CAPTCHAs; using a single type will reduce user errors.