There once was a model quite scenic Many lines but wholly anemic New features requested But nothing was tested The burdened developer feels sick Legacy code is code that makes money, as they saying goes. But we have an opportunity to refactor when tickets and feature requests come in. How are we going to make sure we don't break things? Using a test first approach we will turn this dumb data container into a rich and meaningful domain model. This gives us the confidence to tackle new challenges, without fear of regression. We will create a safe path forward guided by test driven development and new feature requests only. If you know a bit about tests but don't know how to get started in your own code base, then this is the talk for you.


Comments are closed.

Miro Svrtan at 14:29 on 27 Oct 2018

I liked the talk, it was nice to see how others are doing refactoring phases. Speaker did get distracted a few times, and on some points gone very deep while on some went very shallow (towards the end) so it was bit hard to keep track of all the changes.

Through refactoring excerices, I was trying to understand what the code does and it kept bugging me as I'm used to first needing to understand existing code and not go changing it straight away.

For 5*: explain why mocking is the problem :)

P.S. I was sitting in the last row and it was hard to read the code, just an observation

The talk was great and very informative. I liked the flow and the pace, didn't even notice that he couldn't use hist notes. Perhaps the example was a bit too complex, but I think that was the point. He started from not knowing how the code worked at all (and so did we) and by the end of talk it was in a much better shape all done by adding the tests first.

Very good presentation, but font of code on slides was to small and hard to follow

Very nice presentation, a lot of sense of humour.