TestabilityExplorer.org records the testability scores for many open source and commercial Java libraries.

The compiled bytecode for the library is analyzed and metrics are calculated for the testability of individual classes. Those classes fall into one of three categories - 'excellent', 'good' and 'needs work'. Generally speaking, injectability, mockabiliy and composition are good, and static state is bad. Figures are recursively calculated, but only inside the jar in question.

The metrics are a calculation of the skill of the development team in making their classes testable. You cannot use these metrics to say that Tomcat is better than Jetty or vice versa, as the features of each are not taken into account. These metrics will also not tell you whether a particular library will be easy to use or not. It just tells you how dedicated the development team was to making testable software. As we track the changing figures overtime, we can see whether the team in question was dedicated to improvement or not.