http://ayende.comAyende @ RahienCopyright (C) Ayende Rahien 2004 - 2021 (c) 202660I think this is more about what Jeff Palermo meant with good design: http://ferventcoder.com/archive/2008/08/23/test-driven-success-story-plus-side-effects-of-good-design.aspx Good design makes your code maintainable. Testability is only ONE of the benefits. http://ayende.com/3532/inversion-of-control-is-not-about-testability#comment3http://ayende.com/3532/inversion-of-control-is-not-about-testability#comment3Sat, 23 Aug 2008 19:21:04 GMTIf you're going to snipe, at least ****try to understand my point. At the very least, it should be clear from my post that I did ****not miss Polermo's point about Inversion of Control. The fact that testability is beside the point is my problem with how testability nevertheless entwines itself in these discussions. If testability is truly beside the point, then stop injecting it into ****every discussion of IoC/DIP/SoC! Polermo's discussion of coupling would have been ****far more interesting if he would have stuck to the premise of his title. Instead, his asides about testability as a beneficial side-effect distracted from his point and prevented delving further into the things that supported his argument ****better . And I'm not all that fond of your characterization of my motivations, either. Typemock is holy? Please. I beg you to point out anything I've ****ever said that would imply that Typemock is anything more than a useful tool. Freeing us from needing to turn elsewhere for testability is hardly a call for sainthood. Finally, if you're going to use my own posts against me, try to at least represent them accurately (even if you don't want to give me credit for honest and open exploration of a tricky topic). Testing and testability had absolutely nothing to do with the trade-offs explored there and since I didn't use Typemock in the testing of that product (the provider pattern is as easy to test as DI for most test frameworks), that article hardly invalidates anything I've said... ever. http://ayende.com/3532/inversion-of-control-is-not-about-testability#comment2http://ayende.com/3532/inversion-of-control-is-not-about-testability#comment2Mon, 18 Aug 2008 17:09:45 GMTWhat are they talking about anyway? Testing is one thing, design is another. http://ayende.com/3532/inversion-of-control-is-not-about-testability#comment1http://ayende.com/3532/inversion-of-control-is-not-about-testability#comment1Sat, 16 Aug 2008 19:30:00 GMT