Timeline for Does defect density measure QA or dev effectiveness?
Current License: CC BY-SA 4.0
8 events
| when toggle format | what | by | license | comment | |
|---|---|---|---|---|---|
| Jan 15 at 21:21 | comment | added | gnasher729 | I think there was a competition where one group gave 30 reasons why more kLOC means better developers and 30 reasons why it meant less good developers. | |
| Oct 9, 2018 at 9:57 | comment | added | John V | @DavidArno I agree with that. But on the other hand, it is likely that all projects will do that and therefore the distortion will be everywhere. I think that is why the metric was quite heavily used by IBM an others. | |
| Oct 9, 2018 at 9:39 | comment | added | David Arno | @JohnV. There are two problems with using KLOC. Firstly, the more lines I write, the lower the density for a given number of defects. So if you are using defect density to measure my app's quality, it makes sense for me to make my code as verbose as my coding standards allow to make my app appear to be of higher quality. Secondly, unless you have offensively restrictive coding standards that remove all self expression in code, different devs will use different solutions to the same problem. 2-3 times difference in the line count is not unlikely there. So you aren't comparing fairly with KLOC. | |
| Oct 9, 2018 at 6:14 | comment | added | John V | David, if the teams use the same languages, isnt KLOC still more accurate than story points? If there are coding guidelines and conventions, the differente should not be that bad, in my opinion. | |
| Oct 8, 2018 at 15:43 | comment | added | Deduplicator | Yes, lines is a very imperfect measure: While drawn-out code likely has fewer bugs per line (it does little per line, so has little per line to do wrong), it probably has more bugs per unit due to monotony. | |
| Oct 8, 2018 at 14:39 | comment | added | Thomas Owens♦ |
If I simply do not test the code, my defect density is zero: perfect quality! Of course in reality, most untested code is far from good quality. - That is not correct. If you have field reported issues, then you have a defect density of more than 0. Also, not tracking the issues doesn't mean they don't exist.
|
|
| Oct 8, 2018 at 13:37 | comment | added | John V | Yes but I still do not think I measure effectiveness of testing (as I cannot know whether there were 0 issues and hence 0 found, or 100 issues and all discovered and fixed). I think it reflects the maturity of dev-QA. | |
| Oct 8, 2018 at 13:33 | history | answered | David Arno | CC BY-SA 4.0 |