Further Adventures of spam Filtering
A while ago, I redirected my LiveJournal email notifications to my Gmail account. Today I noticed that somebody had commented on a post, but I hadn't received the notification. A quick in:spam livejournal search revealed that not one, or two or three, but four comment notifications had been filed as spam. Eep! Let's hope I can set the right example by setting all four to "not spam" in one swoop.
In other news, I've added another twist to my automated spam trashing. I now have a text file in my home directory that stores timestamps for individual spam arriving at any of the previously established trap addresses. Each receiving address has to have its own .qmail file. Fortunately, programs executed from within those files inherit a set of environment variables that include the filename of the .qmail file. Ding! I just open the log file, lock it, print the time stamp (seconds since the epoch) and that environment variable, then unlock and close the log file. This way I'll be able to see just how awful my spam load really is, instead of the much smaller number I see in my Gmail spam folder. With more than two-hundred addresses being filtered, I'm pretty sure I'm receiving well over a thousand. I also expect to be able to plot things like density vs. time of day. Since the filtered addresses tend to tip me off to the time they were harvested, I might even be able to correlate the harvest time with the spam time. Hmm . . .
In other news, I've added another twist to my automated spam trashing. I now have a text file in my home directory that stores timestamps for individual spam arriving at any of the previously established trap addresses. Each receiving address has to have its own .qmail file. Fortunately, programs executed from within those files inherit a set of environment variables that include the filename of the .qmail file. Ding! I just open the log file, lock it, print the time stamp (seconds since the epoch) and that environment variable, then unlock and close the log file. This way I'll be able to see just how awful my spam load really is, instead of the much smaller number I see in my Gmail spam folder. With more than two-hundred addresses being filtered, I'm pretty sure I'm receiving well over a thousand. I also expect to be able to plot things like density vs. time of day. Since the filtered addresses tend to tip me off to the time they were harvested, I might even be able to correlate the harvest time with the spam time. Hmm . . .