I once argued, in a previous talk, that IR should stand for Interactive Retrieval, rather than Information Retrieval. Unfortunately, the inherently interactive nature of IR has been largely ignored by researchers in the past. It is my belief that IR system evaluation, and consequently also design, is the poorer for this.
Traditionally, IR system evaluation takes the form of a comparison of system performance with subject experts' opinions. This gives rise to a measure of system effectiveness which takes no account of users' judgements.
In this talk, I'll fill in some IR background (no prior knowledge of IR is assumed!), highlight the present deficiencies of IR system evaluation, and examine some ways in which HCI might make a valuable contribution to this process.
For more information about this talk please contact jane@dcs.gla.ac.uk.