An open-source analysis schema for identifying software comprehension processes

This paper presents an analysis schema for talk-aloud protocols, which distinguishes between bottomup and two variants of top-down comprehension (as originally described by Brooks and Soloway et al). The first is ‘expectation-based’ comprehension, where the programmer has pre-generated hypotheses of the code’s meaning before studying the code segments. The second variant is ‘inference-based’ comprehension, where the programmer derives hypotheses from clichéd implementations in the code. The schema is placed in the context of other talk-aloud analysis schemas. It is described in detail, allowing other researchers replicate the protocol and assess the generality of its findings. A brief description of an experiment, which uses this schema, is also included.