tag:blogger.com,1999:blog-6178209.post113073784786640246..comments2007-04-16T05:50:18.553-05:00Igorhttp://www.blogger.com/profile/12832903167357753682noreply@blogger.comBlogger3125tag:blogger.com,1999:blog-6178209.post-1135721968018458612005-12-27T16:19:00.000-06:002005-12-27T16:19:00.000-06:00Hi, I am writing a framework based on BDD concepts. I try to think about 'intents' rather than tests: <A HREF="http://liquiddevelopment.blogspot.com/2005/12/intent-03.html" REL="nofollow">Intent framework</A>Chiaroscurohttp://www.blogger.com/profile/11467489170756712526noreply@blogger.comtag:blogger.com,1999:blog-6178209.post-1131118797345428682005-11-04T09:39:00.000-06:002005-11-04T09:39:00.000-06:00Thanks, for the interest Rob.<BR/><BR/> I will definitely share my experience for phase 5. Actually, our project didn’t finish before phase 5. We had about a month in this phase. We were pretty happy in this phase and didn’t have any substantial problems. We definitely eliminated the problem after refactoring with unnecessary failing tests that use mocks. Also, using BDD was very beneficial for better OOD. However, the time wasn’t enough to be able to make some generalizations about this phase. I couple of things that could be noted are:<BR/> - If TDD seems unnatural in the beginning to most of the developers, BDD is even harder to be apprehend. Since BDD make you much more aware about your design and what you expose, some developer doesn’t feel that you have to go to all these troubles in order to test behavior of an object rather than expose its state. We surely have some reluctance from some of the developers on the team for practicing BDD. However, if most of the developers on a team are convinced in the benefits of using BDD, the result seams to be very impressive. Of course, as everything else, you should do it in the right way in order to have good results.<BR/> - Using stubs could go in the other direction again. So, when we actually need to use mock in order to verify some kind of behavior, we could instead use static stubs. I haven’t seen such a tendency yet but could be a possible outcome.<BR/> <BR/>The time wasn’t enough to be able to do good evaluation on phase 5. However, our new project will be to integrate the application with the rest of the enterprise system. During that time, I will be able to evaluated more fully phase 5 and will share the results.<BR/><BR/>Thanks.Igorhttp://www.blogger.com/profile/12782792313857381015noreply@blogger.comtag:blogger.com,1999:blog-6178209.post-1131043628687419482005-11-03T12:47:00.000-06:002005-11-03T12:47:00.000-06:00Igor, this stuff rings very true to me. I've gone through a similar sequence of learning, and so far I'm at 'phase 4', using mocks even when they are probably over-coupling, resulting in brittle code. I wish you had more information on 'phase 5'. I'd really like to hear more of your experiences in that area.Rob Harwoodhttp://www.blogger.com/profile/03922491243153382760noreply@blogger.com