The agenda suggested that this session would compare the features of a number of competing GUIs, such as Windows, Presentation Manager and X. When I got here, I discovered that the lecture was going to be a rudimentary introduction to GUIs. Too bad.

The speaker was Rosalind Radcliffe, an IBM CUA "User Interface Architect" from Cary, NC. While she knows all about the Windows and OS/2 GUIs, she doesn't know much about the others. She doesn't know much history, either. I disagreed with some of the conclusions she drew about GUIs in general. (Maybe I was just a little cranky because it was Monday morning.)

She talked briefly about cost justification. Implementing GUIs across your network is an expensive task -- Windows and OS/2 are memory intensive beasts, and they like larger processors. She complained that the benefits are all in "soft dollars" that were hard to quantify. Now she claimed that training and support costs go down when you implement GUIs, and I won't dispute this. But I wonder if productivity levels for experienced users actually go DOWN when you put a GUI between the user and most existing character-based applications.

(I guess soft dollars are like pornography; nobody can define or measure it, but everyone knows it when they see it.)

She also talked in nonspecific terms about "increased user satisfaction". I think this is mostly infatuation with a new toy.

Ms. Radcliffe did come up with an interesting observation: not all GUIs are based on the desktop metaphor. Go's Penpoint system (for example) implements a notebook metaphor.