Chapter 9: Applying Support Theories in Field Conditions
The overall aim of this dissertation is to transform tools research
and development by the provision of a theory-based toolkit: methods,
models, and theories. The purpose of this chapter is to explore the
methods possible for applying CoSTH, HASTI, and RODS to engineering
goals. Towards this end, the ways of using the theories in combination
with empirical methods were investigated.
Theories are normally built and verified using various
sorts of empirical studies. When theories are used, it is
often in during analysis (modeling, prediction, explanation, etc.).
However theories can also be used in empirical circumstances.
For instance, a model of program comprehension can be used to code
actions of users so that various processes can be extracted, or
various measurements can be made (e.g., frequency of switching
focus between different hypotheses). In that sense, the theories and
models structure and support empiricism so that questions can be answered
empirically even if the theory or model does not directly answer them
(e.g., frequency of switching focus between different hypotheses).
In this way theories and models can be useful tools to researchers beyond
its particular abilities to predict or explain. It is a tool for asking
interesting questions that can be answered by a study.
This chapter explores methods for using CoSTH, HASTI, and RODS to
pose and answer interesting research and engineering questions in
reasonable tool devleopment scenarios. In particular, ways of
using these resources to explore cognitive support claims
in tools is investigated. The instrument used in this exploration
was a field study of software developers. The tool usage of
several (semi-) professional software developers was studied
and observations were recorded. The participants worked on their
own development and maintenance tasks, using their own professional
tools, and in their normal work environment. The study is
described, and the types of observation and analysis methods tried
are described.
The key part of this chapter is a benchmark analysis of a particular
observation session. A particular subject and session are chosen,
the cognitive support within the tools are analyzed using CoSTH and
HASTI, and a coding scheme is devised for coding the activity. Once
the activity is coded, it is used to drive an exploration of several
types of
|