NotesFAQContact Us
Search Tips
ERIC Number: ED152321
Record Type: RIE
Publication Date: 1978-Feb-1
Pages: 150
Abstractor: N/A
Reference Count: 0
Representing and Teaching Knowledge for Troubleshooting/Debugging. Technical Report No. 292.
Wescourt, Keith T.; Hemphill, Linda
The goal of the present project was to identify the types of knowledge necessary and useful for competent troubleshooting/debugging and to examine how new approaches to formal instruction might influence the attainment of competence by students. The research focused on the role of general strategies in troubleshooting/debugging, and how they might be represented and taught explicitly and directly in order to avoid the cost and other drawbacks of learning indirectly by observation and practice. Related work on troubleshooting/debugging was examined, and in conjunction with a logical analysis, contributed to a characterization of troubleshooting/debugging problems that emphasizes their generality across a number of technical fields and informal contexts. Further data gathered from students learning computer programming suggest that expert debuggers do not necessarily have superior general strategies; rather, their expertise derives from specific and sometimes idiosyncratic knowledge acquired through experience. An attempt to obtain a rigorous characterization of the differences and defects in the debugging strategy of students by applying a model-oriented data analysis method was unsuccessful. Another study was conducted to determine the effects of presenting a tutorial text which describes a few general heuristics designed to correct strategy deficits; results indicated a marginal increase in the apparent use of some of the heuristics by those who studied the text compared to a group who did not. The several methodological limitations and problems encountered suggest that, if the causes of differences in ability are to be specified in detail, and if the effects of direct problem-solving instruction are to be assessed, then it will be necessary to perfect model-based data analysis methods. (Author/DAG)
Publication Type: Reports - Research
Education Level: N/A
Audience: N/A
Language: N/A
Sponsor: Advanced Research Projects Agency (DOD), Washington, DC.; Office of Naval Research, Washington, DC. Personnel and Training Branch.
Authoring Institution: Stanford Univ., CA. Inst. for Mathematical Studies in Social Science.