Actually that’s also an interesting statistic to cover. What’s the proportion of programmers who learnt because they were supported vs unsupported (and while we’re at it do code quality analysis just to see)