2/7/00 abstract
Norman, D. A. (1983). Some observations on mental models. In D. Gentner and A. L. Stevens (Eds.), Mental models (p. 7-14). Hillsdale, NJ: Lawrence Erlbaum.
Discusses mental models. People have incomplete mental models of the machines they use. Models are simpler than the actual objects, omitting details of the operation. Mental models are ``superstitious'', reflecting incorrect understandings or assumptions.
2/7/00 abstract
Rumelhart, D. E. (1984). Schemata and the cognitive system. In Wyer and Srull (Eds.), Handbook of social cognition, vol. 1 (p. 161-188). Hillsdale NJ: Lawrence Erlbaum.
Explains schemata and relates them to human cognition. Schemata are representations of knowledge about objects or events at various levels of abstraction. They are created through experience. A schema for one situation can be used to provide background information in a new, similar situation.
2/7/00 thoughts
Dad called today. ``I can't get my internet to connect, but Dorothy's works fine.'' I tried to step him through debugging the connection, but he said ``I'm already connected.'' After 10 minutes of conversation, I finally determined that there was nothing wrong with his connection to the internet; he was having trouble viewing certain web sites. Dorothy was fine because she was using different web sites. The difference in our mental models is astounding. When I'm teaching, this problem doesn't come up, presumably because the students get used to my mental model of the material. In the past, when I had a computer support job, communication was easier as well. I think the main difference is that we were on the phone rather than being in the same room.
A mental model is incredibly difficult to convey using only words, since there is no way to determine that both parties are using the words in the same way. If the Reddy article had limited communication between the toolmakers to words, they would have never been able to understand each other.
2/7/00 discussion
Wurman, R. S. (1989). The understanding business. In Information anxiety: what to do when information doesn't tell you what you need to know (p. 51-82). New York: Doubleday. 1989.
This chapter starts off by reasurring us that not knowing something isn't the end of the world. Find a way to organize items so that they will be easy to find when you need them.
Wurman then asserts that there are a five ways to organize objects: category, time, location, alphabet, continuum. The continuum class is left open-ended, so in practice, there are still an infinite number of ways to organize things. In fact, time, location, and alphabet are just types of continuums. It would make more sense to say that category and continuum are the only ways to organize things, noting that people often use continuums to organize things within categories, but not vice versa.
Next, we are told that organization simplifies and facilitates generalization. In my view, it simplifies by facilitating generalization.
Wurman spends a lot of time convincing us that it is useful to look for the opposite of any object, category, or classification scheme that we work with. This is an interesting idea, and merits further study. Likewise, I think the ``ask why before how'' section provides a good method for rethinking problems.
2/7/00 thoughts
How to think outside the box (in social science).
2/8/00 assignment
The Shopping Script
actor3 creates list1 (during cooking script)
actor3 supplements list1
actor1 get money1
actor1 goto store1 (using transportation script)
actor1 enter store1
actor1 get cart1
foreach section1 in store1
actor1 enter section1
foreach item1 in (section1.contents and list1.contents)
find item1
put item1 into cart1
actor1 enter checkout1
(execute wait-in-line script)
actor1 give money1 to cashier1
actor1 exit store1
actor1 put cart1.contents into (vehicle from transportation script)
actor1 goto actor1.home (using transportation script)
Even at this level of detail, there are several things missing. Some
stores have conveyor belts where actor1 should place the items. Some
stores (Sam's club) omit this step, but add extra steps (giving
cashier1 an ID card).
General problems with scripts:
2/10/00 thoughts
As I was driving in to class last night, I listened to the Mark Shaw program on WGCL. The discussion was about whether the Boy Scouts of America should be allowed to keep gays and atheists out of their organization. After several arguments had been stated, Mr. Shaw summed up by saying ``It's all definitions,'' indicating that the answer to the question rested on an individual's mental model. Someone's idea of the scope of the Boy Scouts, the government, or gays could make all of the difference in their decision.
2/12/00 abstract
Engle, M. (1998). Remythologizing work: the role of archetypal images in the humanization of librarianship. Available on the web.
Describes stereotypes that are applied to librarians. Discusses how librarians can use knowledge of stereotypes to improve their work.
2/12/00 abstract
DeCandido, G. A. (1999). Bibliographic good vs. evil in Buffy the Vampire Slayer. American Libraries (September), 44-47.
Discusses the character Giles in the television series Buffy the Vampire Slayer. Giles is a dedicated librarian who uses his books to assist Buffy and her friends in their fight against evil.
2/12/00 abstract
Bower, B. (1996). Fighting stereotype stigma. ScinceNewsOnline (june 29, 1996). Available on the web.
Summarizes research of stereotypes. Scientists often shun stereotypes as predudicial, inaccurate, and inflexible. But stereotypes can be helpful representations of the world. Minority groups sometimes emphasize differences between themselves and others to bolster their sense of social identity. Teachers' stereotypes of students reflect actual ability, not differences between minority groups.
2/12/00 abstract
Miller, G. A. (1956). The magical number seven plus or minus two: Some limits on our capacity for processing information. Psychological Review 63, 81-87.
Discusses research that results in human capabilities measuring 7 (plus or minus 2). On many types of one-dimensional judgements, humans can discriminate between 7 classes. As more dimensions are added, more classes can be distinguished. Humans can hold 7 items in their immediate (working) memory, regardless of the size of the items. If several items can be recoded into a composite item, then 7 composite items can be held in working memory.
2/12/00 class thoughts
Jacob Says:
``...in much of AI work, there has been a failure to do their homework [on how the mind works].''
(see discussion in group work section)
Jacob Says:
``They [computers] become intelligent when we can program intention.''
I believe there is a lot more than intention. Modeling goals and plans has been a focus of AI from the start, and it's fairly well understood. The big problem is knowledge. An AI system doesn't know the consequences (or non-consequences) of its actions unless it has an explicit representation of all the possible consequences as well as adequate ways of detecting them. The state of the world can also change regardless of the computer's actions. This makes it very difficult to perform subsequent actions. Perhaps it would be better to say that computers become intelligent when they can modify their intentions based on the changing state of the world.
2/12/00 group work
Our group assignment for this week was to create a shopping script, using the scripts we had created individually as a basis. Group 2 came up with a script that was more abstract than my individual script. The abstract script is better because it can deal with more types of situations, but it is too abstract to be of real use in guiding the actions of a computer/robot.
2/12/00 thoughts
In the discussion of group work, Dr. Jacob asserted that scripts are too restrictive and can't deal with the unexpected. Yes, but they are particularly useful in certain situations. People tend to perform tasks in set ways. Microsoft Office uses ``wizards'' as a way to let people work through a script for a particular task. Of course, the user must select the script before it is used. Current work at Northwestern's Intelligent Information Laboratory uses scripts to detect common actions that people perform when they are giving lectures and browsing the web. These scripts are then tied to AI systems that assit users by anticipating their next action.
2/14/00 thoughts
Categorization explained
We have been told that humans need to name things, and the name they give to a thing is usually based on its functional usage (Brown).
Miller's article from last week's recommended reading tells us that our working memory can hold 7 (plus or minus 2) items, regardless of the complexity of the items. Given this limit, it seems necessary to categorize things in order to work with them.