You're comparing different things - one is current levels of CompSci (small, limited, sample size - hell I wouldn't show up in CompSci numbers) which is also ba

You're comparing different things - one is current levels of CompSci (small, limited, sample size - hell I wouldn't show up in CompSci numbers) which is also basically at an all time high. The other is a total size metric that won't show churn. I get what you're trying to show but the comparison doesn't follow.