Unit Testing, Agile Development, Leadership & .NET - By Roy Osherove
Ads Via DevMavens
I’ve always been looking for real research numbers on the effectiveness of TDD. For future reference, there are several listed here:
http://biblio.gdinwiddie.com/biblio/StudiesOfTestDrivenDevelopment
Оценка эффективности TDD в исследовании National Research Council Canada на тему
Dear Roy!
Since you have been to Norway a couple of times now, I would like to inform you of a norwegian research laboratory - Simula.
They have some interesting publications:
simula.no/.../publications
Alexander, this is an English blog...so why are you posting in Russian?
Interesting but to me, not compelling.
My experience with TDD has been so far dismal. It has not improved the quality of my code (although I step through all of my new code, and my quality is pretty high to begin with), has not caught errors, and has made fixing bugs actually more difficult in several cases.
So is my experience unusual? Perhaps, perhaps not. TDD may be useful for new programmers, and it may be useful for managers. For someone like me who can easily make software from requirements and tests as he goes by stepping through new code, TDD is a burden.
What's more, what is the cost on a TDD project when the requirements change? How much pushback will you get from engineers who may have to rewrite hundreds of tests?
Is TDD right for everyone? What about the guy in the corner who does the work of ten (or more) and happily codes until dawn? Every organization has at least one of these guys, and I've never met one who was dogmatic about tests. Do you really want to kill their productivity and/or spirit with TDD?