January 31, 2004
Statistics
As some of you may know, I graph statistics on all sorts of things.
This morning I discovered that my temperature graphs don't know what to do when it gets below zero. Which is just as well, since I don't either.
I've added a new measure to my stats page. Up until now, I've been graphing email/spam as a cumulative for the month, which completely fails to reflect the seriousness of the latest flood of spam.
Although, on the graph, you can very clearly identify where the flood started, and although the spam percentage gradually creeps up, it has an entire month worth of 62% spam that is has to offset to make any changes. However, from watching the logs, I felt that spam levels were much closer 90%.
Turns out I was right. In the new graphs, you can seldom see any daylight between the spam graph and the total email graph, and spam levels are right around 91%. At my peak, I'm processing about 60 email messages in 5 minutes, but only 3 of those made it into anyone's inbox.
Given that email is such an important part of what the internet carries, 90+% of it being essentially stolen resources is pretty serious. I'm hoping that the courts will authorize public floggings of spammers and virus authors.