The HCI Research Centre
School of Computing and Mathematics
University of Huddersfield, Huddersfield, UK.
email: A.J.Dix@soc.staffs.ac.uk
Invited talk at the BCS FACS Christmas Workshop
Imperial College, London on 19th December 1995
Program language semantics are, in common with all formal semantics, by definition meaningless. A formal semantics can at best give meaning relative to some context. A formal specification of a sort algorithm has no meaning until it is associated with a real programming language, but of course the language itself only has meaning with respect to a compiler etc. ...
User interfaces are to some extent different. We have a context given to us - the human user. Formal semantics for user interfaces do therefore have natural boundaries: the user input on one side and system display on the other. We can build formal models of this nature and describe some interface usability properties over the models.
Unfortunately even that does not ground the semantics entirely. To really capture the meaning, we need to understand human perception and cognition. But people are so wonderfully unpredictable ... are we on a highway to nowhere, trying to formalise people?
For modelling cooperative work the situation is more extreme. Not only do we have individual psychology, but also social processes at work. Again, one approach is to capture these formally, but perhaps that is not necessary. In cooperative work, the critical things are not so much what people are thinking, but the external representations that they use and their interactions with one another. In some ways formalising cooperative systems may be easier than those for individual users.