Archive for May 21st, 2010

Psychoanalysis of Software Troubleshooting and Debugging (Part 1)

Friday, May 21st, 2010

In this part I would like to introduce the notion of Forgotten Facts in opposition to Basic Facts or supporting information. These are facts that engineers often feel uncomfortable to mention because they are troubleshooting information they couldn’t obtain (if they tried) due to some time or customer pressures, failures, incorrectly understood troubleshooting procedures or some other obstacles. Therefore it is important to have a set of counter questions or checklists mapped from common software behaviour patterns to software troubleshooting patterns. Problem descriptions should also be subjected to close reading to reveal unconsciously concealed information. Next part will explore this in more detail with some case studies commissioned by Software Maintenance Institute.

- Dmitry Vostokov @ DumpAnalysis.org + TraceAnalysis.org -

Two Readings of a Software Trace

Friday, May 21st, 2010

When we have a software trace we read it in two directions. The first one is to deconstruct it into a linear ordered source code based on PLOT fragments. The second direction is to construct an interpretation that serve as an explanation for reported software behaviour. During the interpretive reading we remove irrelevant information, compress relevant activity regions and construct the new fictional software trace based on discovered patterns and our problem description.

- Dmitry Vostokov @ DumpAnalysis.org + TraceAnalysis.org -