|By Shamus||Jun 15, 2014||175 comments|
We have a problem. The problem is so old and so commonplace that we’ve all gotten used to it. But it’s still a problem. The problem is that the WordPress moderation filters are comically primitive. They’re not even up to 1997 email-filtering standards. In fact, I’m starting to suspect that the spam filter is just a random number generator that marks every 20th comment as spam.
On the top we have Henson, who has posted a small comment that contains no common spam keywords. This was posted to the most recent episode of Spoiler Warning. It contains no links. Moreover, Henson has successfully left 64 comments in the past without being flagged as a spammer.
On the bottom we have “residential steam showers”. It’s also worth noting that:
- This “person” has never commented before.
- This comment was left on a post that is half a decade old.
- It is loaded with spam phrases that I have marked as spam again and again and again. (What is with you spammers selling showers and bathroom fixtures? Even if I left every single comment stand, your spam would NEVER build up enough search engine credibility to end up anywhere NEAR the top of the search results. It will never happen. Give up.)
- It features a long gibberish URL, which is a common trait among spammers.
But Henson was inexplicably marked as spam, and not residential steam showers. Then we have this:
ps238principal has successfully left ONE THOUSAND SIX HUNDRED AND THIRTY-EIGHT non-spam comments. Yet the spam filter felt the need to flag this reasonable, inoffensive comment as spam.
On the top, “real estate” is leaving a word-for-word reproduction of a comment I’ve marked as spam a hundred times in the past. On the bottom, “hack les simpson” is leaving a comment with goofy manual line breaks that are common to 80% of all spam and is never done by any human ever. They’re also loaded with phrases that are very common to spammers. (Seriously, spammers love to tell me how nice my site looks. Also they love to use the word “fastidious”. Incorrectly. As in, “this post is in fact a fastidious one it helps new \n net visitors, who are wishing for blogging”.)
I guess it flagged ET because the comment had two links? But they’re to youtube, just like the spam below it. And ET has nearly 700 valid comments. “Doctor Oz” has zero, plus goofy line breaks and spammy content. And for the record, the “1 comment approved” means THIS comment. It doesn’t mean I’ve approved a comment from them in the past.
Also, I am reminded how nice it is to be rid of the Google Adbot. I can write all this without worrying about pissing it off or being paranoid about what ads it will choose based on my content.
OBAMA DOESN’T WANT YOU TO KNOW THIS TRICK FOR ONE CLICK UNDERAGE PAYDAY LOANS FOR FAST WEIGHT LOSS SHOWER HEADS, NO PRESCRIPTION REQUIRED!
Tell me again about the great strides we’re making in artificial intelligence. </facepalm>
This is beyond pathetic. If you can’t recognize these three flagrantly obvious spam comments as spam, then you have not written a spam filter. I don’t know what your software is doing, but it sure as hell isn’t looking for spam. Once again: Steam showers and sex toys with goofy line breaks and sketchy URLs on ancient posts.
This is the tyranny we live under. Our spam filter is like an airport security checkpoint that waves through men in sunglasses with ticking briefcases that have giant nuclear symbols beside digital countdown timers. But then the guards body tackle and strip search little old ladiesSo, not all that different from real airport security, really.. It would be one thing if it looked like a slightly buggy system that missed every once in a while, but this is so bad I can’t even tell what it’s using as criteria for spam.
Even more embarrassing: This circus of failure is actually the result of three spam filters: Akismet, GROWMAP, and Bad Behavior.
And to be completely fair: Yes, they do catch more than they let through. My comments would by 90% spam without them. Also, Growmap doesn’t do filtering based on content. It just puts the “Confirm you are not a spammer” checkbox in there. So it really just cuts down on the volume of crap the other two have to cope withWhen I first installed it, Growmap worked like magic. No spam for weeks. But spammers always adapt..
I could tolerate the occasional spam getting through. But what I can’t fathom are these false positives. There is no pattern or reason to them.
So if you’re curious why sometimes your harmless comment was put into moderation, now you know: NO REASON WHATSOEVER.