Readings Nov 8 1994

Christian mogensen@cs.stanford.edu -=- http://www-pcd.stanford.edu/mogens/377/reaction-1108.html

Could you please list the titles of the readings so that we don't have to guess which ones are included and which are not? Thanks.

Human Information Processing

The idea of "refresh memory" (in computer terms) or "sensory information storage" as they refer to it is great. I note that refresh rate is roughly one to two Human Information Processor perceptual cycles (100-500 ms).
Since data is being retrieved at a much faster rate (50 ms or less), this implies that the SIS is either a multiway cache or it is doing averaging on the sensory data streams, combining inputs with data from a few cycles ago.
I want to know more about the control processes so briefly mentioned in the text. Looks like I ahve to find Ch. Nine.
I enjoyed the examples and mini experiments. Perception: fun for the entire classroom.
Ok, so language and dinner last Sunday (pesto) are both in long term storage. Obviously we don't access them in the same way. This would be interesting to investigate a bit further, but presumably all is revealed in Ch. 10. Argh! Frustrating. They say it's the most difficult problem (retrieval) but then say nothing more about it.
Language is a can of worms in its own right: Chomsky's language processor and the general idea of human mind as symbol manipulator crop up here.

The phrase "remembering future events" is a great hook for a science fiction story. It's probably been done to death by now though.
It was intriguing to see that the authors consider their work in the larger context: the social costs of memory failures and the differences between them; the roles of commitment and motivation.
The idea of external vs internal memory aids ties nicely in with the idea of cognitive load and costs - it is easier to read a note than to remember the text of the note. i.e. moving from cognition to perception.
The summary at the end of this section is golden - very useful and summarizes all the sections (including the ones not in the reader).

Speech is a social activity - reading is a solitary act, and engages us in different ways. They mention this in passing but do not elaborate on the differences - it's nice that they mention it, but it seems as if the only thing of interest (judging from the experimental work cited) that no work is being done comparing speaking to reading. Another interesting thing to look at would be speech in a restricted domain, like telephone speech - how is speech affected by the lack of other cues, etc...
Speech recall from the highly condensed form stored in long-term memory is not likely to match what was actually said except for themes and general structure. Recall is colored by motives and wishes of the person recalling. This implies how speech interfaces are used will affect users' recall. (Note also Nass: "Computers as Social Actors")
e.g. if the user is told about errors in a personal way ("you", implied causation) then the user will remember the error, but also resent the computer for pointing this out. Result: an efficient worker with angry feelings towards the terminal.
The chapter is nicely laid out, with boxed info off to the side. My only wish, since they provide so much background info, would be that they include the experimental descriptions but separate them from the main discussion (or vice versa, there seems to be a lot of experimental data). This would make it easier to refer back to, since the key points being made could easily be identified.