Skip to content

Assessing initial smells

Nic McPhee edited this page Apr 7, 2014 · 5 revisions

Since we don't know this code base, I would recommend we start by letting tools and metrics guide us to areas in need of attention. Examples include (but definitely aren't limited to):

Before class (Thursday, 10 Apr) I'd like everyone to:

  • Set up Maven.
  • Check out my fork (or fork that and check out your fork – I don't really care which).
  • Get the project to build.
  • Spend an hour or so going through warnings from the Google CodePro plugin or (if you have trouble getting it installed and working) Simian, PMD, or Eclipse warnings.
    • If your birthday falls on an odd day of the month, focus on duplicate code.
    • If your birthday falls on an even day of the month, focus on long/complex methods/classes.
    • It's most important that you spend time going over some sort of warnings, so don't get massively stuck if you can't get a given tool or plugin (like CodePro) to work. If you can't find someone to give you a hand, just move on to another tool including, in a pinch, turning on all the Eclipse warnings (except for the generics warnings, which I think are mostly noise) and seeing what that turns up.

Create a page over on the University wiki (not here on the Github wiki). On that page list 5 specific smells you'd be potentially interested in addressing, or that you think the class as a whole should consider addressing. Don't just take the first five things on the list spit out by your tool. There are dozens and dozens of substantial code duplications, for example, and some are definitely more interesting than others. If everyone takes the first five that come out of Simian or CodePro then we'll have a pretty boring conversation on Thursday.

Clone this wiki locally