Often in class, there has been discussion of how much a user of GIS needs to know about the GIScience behind a particular concept or application of GIS. The HCI article by Haklay and Tobon bring this issue to the forefront once more in trying to understand how people interact with computers and GIS in particular.
My impression of this paper and the topic of HCI in general is that experiments, workshops, discussions etc. focused primarily on understanding the difficulties users encountered in working with a GIS and then adapting the software and interface to better cater to the needs of the people. Although this can certainly be helpful for users if the required improvements can be incorporated into a GIS, it does not teach the users the GIScience behind the GIS leaving them vulnerable to making assumptions or conclusions without considering such issues as the uncertainty that could be present in their final product or query response. Teaching the users of these GIScience issues falls under HCI as the method of teaching during an experiment or workshop would likely be very influential on how well participants navigated and used the GIS. As such teaching could be an important area for additional research in HCI.
Catering to different learning styles was not greatly mentioned in the paper although I think this would be another way to improve HCI. GUIs are very helpful in this regard as they cater to visual learners and to some extend those who learn best by doing and trying things as what steps or uses a tool has if it can be visualized. However, those who require explanation or a demonstration to learn may be disadvantaged by GUIs. Tutorials and video demonstrations could be incorporated into GISs to explain how to accomplish particular tasks. Incorporating various learning styles into the GIS would assist users with self-help and reduce, but likely not eliminate, the need for face to face explanation.