My hairiest bug war stories
Marc Eisenstadt · Communications of the ACM · 1997
The TrawlIn early 1992, I posted a request for debugging anecdotes on the BIX electronic bulletin board and followed with similar messages posted to AppleLink, CompuServe, various Usenet newsgroups, and the Open University's own conferencing system.The original message is shown in Figure 1.The trawl request elicited replies from 78 "informants," mostly in the U.S. and the U.K., including implementors of very well known commercial C compilers, members of the ANSI C++ definition group, and other commercial software developers.A total of 110 messages were generated by 78 different informants.Of these, 50 specifically told a story about nasty bugs.A few informants provided many anecdotes; a total of 59 bug anecdotes were collected.Figure 2 shows some typical replies to the original request.The full set of replies and analyses is available from the author. Analyzing the AnecdotesOur analysis included three dimensions: "why difficult," "how found," and "root cause."Although the root cause of reported bugs is of a priori interest, in order to fully characterize the phenomenology of the debugging experiences, I looked at more than the causes of the bugs.After summarizing the data several times, it became apparent that it would be nec-