2.2 Factors affecting defect "reachability" and "triggering"

Mib Software
High Reuse Software
Development and Consulting
We extend and customize
reusable and open software.

Some Implications of Bazaar Size

Third Draft. Aug 11, 1998
Copyright 1997-1998, Forrest J. Cavalier, III Mib Software
All rights reserved. Comments welcome!

2.2 Factors affecting defect "reachability" and "triggering"

A defect remains hidden until it is installed and that code used.

Unlike the social factors affecting the installed effective size, more technically analytical factors affect the likelihood of a defect being "reachable" (compile-time and theoretical conditions) and "triggered" (run-time conditions.)

Since this step is sequential with defect installation, it is important to note that the effective working size for defect reachability is not larger than the number of sites where the defect is installed.

Factors which decrease reachability
It is common to have compile-time conditional sections of code. Depending on the local configuration and installation settings, some sections of code may be completely omitted. Other sections will be "dead" - having no valid execution path and conditions which cause them to be reached, even theoretically.

Factors which decrease code triggering
Many sections of code are run very infrequently, perhaps on low memory conditions, or disk full limits. Others are run only with a very uncommon local run-time configuration or parameter, such as a certain entry or pattern in a configuration file.

For example, most of a program will be present and reachable, perhaps 80%. Perhaps 50% of that 80%, will be run in almost all installations. But the other sections of code (which may actually be a majority) will only run in a small subset of the installed user base, if at all. The defects in these sections of code will have a very much lower chance of being triggered.

Strategies to counter low effective size

Recruit users who requested or will use the new features and pester them to upgrade.
Rely only on user-developers to test.
Alternate defect detection strategies (See 3. Alternate Debugging methods) including the use of automated test cases.

Next: 2.3 Factors affecting defect "detection"

Up to: 2. Debugging by testing dramatically lowers effective size
Up to: Some Implications of Bazaar Size