As I’ve mentioned before, I have modified the Movable Type code to block spammers’ comments, at which I’m reasonably successful (to judge by the record of failed attempts in the activity log). The conditions of use of this software don’t permit me to share actual code, but I’m willing to explain privately the principles I’ve used to anyone who wants to do the same. Anyone with a reasonable grasp of Perl would be able to replicate what I’ve done or do something equivalent.
\nAny spam trap has the problem of ‘false positives’, i.e. what to do about genuine comments which get misclassified as spam. The response to messages which fall into the trap is sufficiently neutral not to offend a genuine contributor and I have access to the text of the comment so that I can check whether it is spam or not (in fact no false positives of this kind have occurred to date).
\nI won’t be giving away too much if I say that some of the spam is identified because it uses a word which occurs on a blacklist. But here I have to be careful setting up the blacklist because these words could conceivably turn up in a genuine posting or a comment on it. I can’t ban the word ‘casino’, for example, because it might occur in a review of a performance of Lulu<\/i> (or indeed one of the various other operas which feature gambling). ‘Bankruptcy’, another favourite spam topic, falls foul of the same opera. Ban ‘cigarette’ and you’re restricted in what you can say about Carmen<\/i>. The root of the problem is that opera and spam share a fascination with human vices!<\/p>\n
[This entry was – perhaps inevitably – targeted by spammers. I’m closing it to comments but if you wish to comment, you may email me and I’ll edit it in.]<\/p>\n","protected":false},"excerpt":{"rendered":"
As I’ve mentioned before, I have modified the Movable Type code to block spammers’ comments, at which I’m reasonably successful (to judge by the record of failed attempts in the activity log). The conditions of use of this software don’t … Continue reading