Computer tools for cognitively challenging activities are considered
useful, to a great extent, because of the support that they provide for
human thinking and problem solving. To analyze, specify, and design
cognitive support, a suitable analytic framework is required. Theories of
"distributed cognition" have been offered as potentially suitable
frameworks, but they have generally failed to plainly articulate
comprehensive theories of cognitive support. This paper seeks to clarify
the intellectual foundations for studying and designing cognitive support,
and aims to put them in a form suitable for design. A framework called RODS
is described as a type of minimal, lightweight intellectual toolkit. Its
main aim is to allow analysts to think in high-level cognition-support
terms rather than be overwhelmed by task- and technology-specific
implementation details. Framing usefulness in terms of cognitive support
makes it possible to define abstract patterns of what makes tools
"good". Implications are drawn for how the framework may be used for
the design of tools in cognitively challenging work domains.