Comments by "" (@ContinuousDelivery) on "Refactoring Legacy Code STEP BY STEP (Part 3)" video.
-
Emily thanks for the great feedback.
The point of my video series was to try and demonstrate the uncertainty that is inherent in this kind of exercise. So I didn't rehearse it or plan it. I am not sure that I was clear enough on the trailing comma thing. I should have been more explicit that at that moment I was experimenting to see how the code worked before I decided how to proceed. This would have been clearer, perhaps, if I hadn't take the short-cut of not really committing to save time. There is no way that I would have committed the change with the line of code commented out!
Yes, I hummed and harr'ed about showing the creation of the Approval test - I think that I will do another separate video on that sometime. I basically took the snippet of sample XML that was in the comments in the code, wrote a test based on that as the input, then measured coverage and added more XML that I guessed would increase the coverage until it did.
Whether or not the code counts as "Testable" based only on the Approval tests is debatable I guess. Strictly I guess you are correct, but I suppose that I fall into the trap of the overloaded nature of the word "Test" in the context of TDD. What I really mean by "Testability" is "Designable Through Executable Specifications". Approval tests don't do that, which is why I reacted against them a bit when I first heard of them from you. I know see their value, but it is not the same thing that you get from TDD. I suppose that by sports analogy, Approval Tests are defensive tests and TDD tests are Offensive tests?
You are probably right that the video would have been in better context if I had a planned feature to add.
Anyway, thanks again for the feedback.
12
-
2
-
2
-
1
-
1
-
1
-
1