Difference between revisions of "Tracking Code Quality"
From CitconWiki
Jump to navigationJump to searchLine 2: | Line 2: | ||
* Presentation by Marty Andrews | * Presentation by Marty Andrews | ||
* The [http://en.wikipedia.org/wiki/Cyclomatic_complexity Cyclomatic Complexity] of a method is really the minimum number of unit tests needed to test very path in a method. | * The [http://en.wikipedia.org/wiki/Cyclomatic_complexity Cyclomatic Complexity] of a method is really the minimum number of unit tests needed to test very path in a method. | ||
− | * By reducing CC, usually by breaking large methods down, you reduce the number of tests you need to write ! | + | * By reducing CC, usually by breaking large methods down, you reduce the number of tests you need to write! |
* Complexian - www.cogentconsulting.com.au/resources/complexian/index.html | * Complexian - www.cogentconsulting.com.au/resources/complexian/index.html | ||
* FindBugs - usually needs human to interpret. Good to hook in but not run in CI. Human can run so often | * FindBugs - usually needs human to interpret. Good to hook in but not run in CI. Human can run so often | ||
Line 12: | Line 12: | ||
** Custom Columns - Define your own quality metric by combining any of Clover's built in ones | ** Custom Columns - Define your own quality metric by combining any of Clover's built in ones | ||
** Track code quality over time using historical charts and reports | ** Track code quality over time using historical charts and reports | ||
+ | |||
+ | === Industry standard CCN numbers === | ||
+ | * There was some discussion over what is an unacceptably high CCN. I ([JerryShea]) don't like to see CCN above 20 personally. I have seen this table reproduced in a few places, including [http://www.linuxjournal.com/article/8035 this linuxjournal article]: | ||
+ | {| class="wikitable" | ||
+ | |- | ||
+ | ! CCN | ||
+ | ! Interpretation | ||
+ | |- | ||
+ | | 1-10 | ||
+ | | a simple program, without much risk | ||
+ | |- | ||
+ | | 11-20 | ||
+ | | more complex, moderate risk | ||
+ | |- | ||
+ | | 21-50 | ||
+ | | complex, high risk | ||
+ | |- | ||
+ | | 51+ | ||
+ | | untestable, very high risk | ||
+ | |} | ||
=== References === | === References === |
Revision as of 13:52, 29 June 2008
Tracking Code Quality
- Presentation by Marty Andrews
- The Cyclomatic Complexity of a method is really the minimum number of unit tests needed to test very path in a method.
- By reducing CC, usually by breaking large methods down, you reduce the number of tests you need to write!
- Complexian - www.cogentconsulting.com.au/resources/complexian/index.html
- FindBugs - usually needs human to interpret. Good to hook in but not run in CI. Human can run so often
- SourceMonitor
- JOODI - Package level checking. Good to enforce certain layers only access other layers e.g.
- Spec#
- Clover2
- Project Risk Coverage Cloud - A tag cloud containing every class that exposes those which have low test coverage but high cyclomatic complexity
- Custom Columns - Define your own quality metric by combining any of Clover's built in ones
- Track code quality over time using historical charts and reports
Industry standard CCN numbers
- There was some discussion over what is an unacceptably high CCN. I ([JerryShea]) don't like to see CCN above 20 personally. I have seen this table reproduced in a few places, including this linuxjournal article:
CCN | Interpretation |
---|---|
1-10 | a simple program, without much risk |
11-20 | more complex, moderate risk |
21-50 | complex, high risk |
51+ | untestable, very high risk |
References
- T.J. McCabe, "A complexity measure," IEEE Trans. Software Eng. vol. SE2, no 4, pp 308-320, 1976
- B.A. Nejmeh, "NPATH: a measure of execution path complexity and its applications," Commun, ACM, vol 31, no 2, pp 188-200, 1988.
- Wikipedia:Cyclomatic_complexity
- Complexity Analyser for Java, http://www.martyandrews.net/resources/complexian.html
- Complexity Analyser for C++, http://gnocchi.sourceforge.net/
- Clover2, http://atlassian.com/clover