Hey there! jeremydmiller is using Twitter.

Twitter is a free service that lets you keep in touch with people through the exchange of quick, frequent answers to one simple question: What's happening? Join today to start receiving jeremydmiller's tweets.

Already using Twitter
from your phone? Click here.

jeremydmiller

@sbohlen Sometimes you do need to make the refactoring first just to enable the testing. Cutting a seam first is often the way to go about 6 hours ago from TweetDeck in reply to sbohlen
Dreary outside today. I think I'm switching to more upbeat music soon. about 6 hours ago from TweetDeck
@sbohlen To me, all it comes down to is that its special abilities aren't nearly as useful as the TypeMock advocates claim. It's not evil. about 6 hours ago from TweetDeck in reply to sbohlen
@sbohlen ...are IMHO just testing the accident of its implementation about 6 hours ago from TweetDeck in reply to sbohlen
@sbohlen Or put it this way, tests, and especially characterization tests, need to be goal oriented. Mocking threads of a legacy furball... about 6 hours ago from TweetDeck in reply to sbohlen
@sbohlen Do you really think someone with less experience can use mock objects in a way that doesn't shoot themselves in the foot? about 6 hours ago from TweetDeck in reply to sbohlen
@sbohlen ...outcomes like "what's on the screen" and "what got into the DB?" 'sides, there's rarely a good seam in the furball to use. about 6 hours ago from TweetDeck in reply to sbohlen
@sbohlen I still don't agree. With legacy code I've always found that you need to start outside in with bigger tests that measure... about 6 hours ago from TweetDeck in reply to sbohlen
@sbohlen ...they'll naturally get discouraged and quit. I truly don't believe that unit testing w/o consideration for design can succeed about 6 hours ago from TweetDeck in reply to sbohlen
@sbohlen Now you're getting into my old argument with Roy. I don't believe that. I think if they try to test really nasty coupled code... about 6 hours ago from TweetDeck in reply to sbohlen
@kevindente ...little control you have over the page construction process about 6 hours ago from TweetDeck in reply to kevindente
@kevindente MVP works with WinForms/WPF/SL because they are stateful. It wasn't the pattern so much as the state mgmt and how... about 6 hours ago from TweetDeck in reply to kevindente
@kevindente The problem w/ MVP and WebForms is the leaky abstraction issue w/ WebForms only pretending to be stateful about 6 hours ago from TweetDeck in reply to kevindente
@sbohlen I'm arguing that even if TypeMock were free, mocking SharePoint API's directly might do more harm than good about 6 hours ago from TweetDeck in reply to sbohlen
@sbohlen Isolate the ugly. Wrap thin gateways around SP. Put the behavioral logic in Presenters. about 6 hours ago from TweetDeck in reply to sbohlen
@sbohlen My old mentor at Dell had a saying for this type of scenario: "I have complete confidence in your ability to manage the situation" about 6 hours ago from TweetDeck in reply to sbohlen
@sbohlen I will grant you SL though, but maybe not for the reasons you think. You cannot simulate a lot of control events w/o TypeMock about 6 hours ago from TweetDeck in reply to sbohlen
@sbohlen Sharepoint is not my problem;-) -- and I'm dubious that mocking SP API's is a good idea anyway. about 6 hours ago from TweetDeck in reply to sbohlen
@sbohlen Silverlight you beat by testing the hell out of the Presentation Models and hoping it's good enough. MS screwed us on that one. about 6 hours ago from TweetDeck in reply to sbohlen
@sbohlen ....a good idea to mock the interaction to a component/technology outside your own codebase -- even tho TypeMock will let you do it about 7 hours ago from TweetDeck in reply to sbohlen