tag:blogger.com,1999:blog-1664217632350055880.post1713026628293781613..comments2009-11-30T13:50:27.172-07:00Michael Murrayhttps://profiles.google.com/109073345657761090655noreply@blogger.comBlogger2125tag:blogger.com,1999:blog-1664217632350055880.post-51994969907942521192009-11-30T13:50:27.172-07:002009-11-30T13:50:27.172-07:00Yes, this is a good start for a great idea (code kata done slowly).<br /><br />Right now, I&#39;m struggling with the same kata and I have a problem looking at your AnagramsFinder and IAnagramGrouper. The result of both is the same and the tests (apart from setting up a not needed IFileParser for the AnagramsFinder) will be the same too.<br /><br />I think this is because you&#39;re using the pipes and filters-pattern (opposite to David, who combines his both results in &quot;focus on test driving logic&quot;). Testing just the filters should be sufficient here?<br /><br />What you maybe can do is a behavioural test. So, instead of testing with stubs, use mock-objects testing weather the methods were called.<br /><br />What do you think about that?<br /><br />p.s. it&#39;s hard to post a comment here because the word verification-textbox is only accessible when selecting text with the mouse and scrolling down :(Tobias Walterhttp://www.blogger.com/profile/00608224689808794198noreply@blogger.comtag:blogger.com,1999:blog-1664217632350055880.post-75630787995421473332009-11-17T18:56:57.211-07:002009-11-17T18:56:57.211-07:00Looks good Mike!<br /><br />I wouldn&#39;t call the previous spec a mistake or &quot;foolish&quot; though. I seen working on simple cases like Count advocated lots of places for doing TDD, I&#39;ve just always struggled with getting from those cases to the main responsibilities of the class.<br /><br />Looking forward to the rest of the series.Davidhttp://www.blogger.com/profile/05155410712205848106noreply@blogger.com