Weird TDD report
Someone pointed out this interesting research report from Microsoft Research about TDD.
The research report claims that TDD results in higher quality, but that it also results in an increase in development time. This sounds rather surprising and weird to me.
When reading the article, however, a few lines are kinda weird. They seem to suggest that they use TDD, the microsoft way. This means, write all test cases and then implement all of them, which is a huge misunderstanding of what TDD is and how it works.
Interesting sentences:
- "The projects also took 15% extra upfront time for writing the tests"
- Figure 3 is interesting... certainly not a continuous integration environment
- In the conclusion: "while the development of both systems utilizing TDD took extra time upfront"
These sentences make the report very uncredible.
0 TrackBacks
Listed below are links to blogs that reference this entry: Weird TDD report.
TrackBack URL for this entry: http://www.odd-e.com/cgi-sys/cgiwrap/basv/managed-mt/mt-tb.cgi/693
Up to my understanding they also did not measure the maintenance time/costs. They did measure "defect density", but neither amount of time required to fix these defects (especially if they are uncovered months after development), not what could be a potential loss/income. Those the latter item certainly depends on the concrete business conditions :)