Consider cyclomatic complexity as part of the health score.

[isoos]

An implementation of such analysis:
https://github.com/daniel-v/codemetrics

[pq]

cc @bwilkerson who might have some thoughts based on experience implementing similar metrics for Java back in the day...

[jonasfj]

Note. I like this idea, but we have plenty of metrics we score on, let's make those better and explain them better before we add opaque metrics like this.

[eernstg]

A good measure might be

weighted_average_of_complexity_metrics / inherent_complexity_of_problem

It wouldn't be particularly helpful to systematically punish every package that solves a complex problem, and it might not be possible to write solutions to complex problems with low complexity metrics. Now we just need a really robust and reliable way to measure the inherent complexity of the problem. ;-)

[sigurdm]

Personally I think the complexity metrics are cool, but way to subjective to start scoring packages on.

Maybe we can use it in other ways.

[jonasfj]

we're not planning to do this.