Remember the days before SPAM? Well you can't because SPAM is not only still here, but is worst every day, and not only affects email but also is having a huge negative effect on weblogs.

Automattic Kismet or Akismet, has become the SPAM tool of choice for the blogosphere. The tool is very effective to control SPAM. To leverage the common knowledge learned from individual weblogs, the tool uses a central knowledgebase or blacklist to quickly share SPAM offenders with every weblog that uses Akismet.

The central repository or blacklist sounds logical, and is a great idea… until you get classified as a false positive.


“A false positive occurs when the system classifies an action as anomalous (a possible intrusion) when it is a legitimate action.”

And when you get classified as a false positive, as it has occurred to me, if really affects one of the most important elements or concepts that fuel weblogs — community participation.

For more than a month, almost NONE of my comments that I've left on weblogs from Mike R, to Tom L, to Ajit, to Om, to MobHappy, to many others, just don't go through.

Even though I have contacted some of the blog editors/owners, and they have addressed the problem, the problem persists. The retraining doesn't work I guess? There is just too much SPAM and it is just a hassle for weblog editors to try to reclassify false positives. SPAM is sucking or will suck the life out of weblog comments (and comments are what makes weblogs interactive), same as it is sucking the life out of Email, as Om recently wrote. (As a side note, I just left a comment @ Om's saying this about the fate of weblog comments, and the comment didn't go through! heh, maybe it needs to be approved first, but I bet you a nice cold dark beer that it has been bit-bucket, same as before).

Akismet is great with SPAM, but as with any rules-based or self-learning solution, it will create false positives and needs a good method for retraining; the current method is too time consuming or too much hassle, or maybe there is just nothing that can be done about it as the SPAM volume is just too great. This is the reason why I don't use Akismet on my blog and instead rely on a simple math question, which is not perfect, but has been good enough (knock on wood) so far. Note that because of the way Akismet works, if someone unintentionally (or not) classifies a commenter as SPAM, that commenter is effectively kept out of the blogosphere from that point on (since Akismet is deployed in so many weblogs).

Akismet is for better; the Akismet website reports that that 93% of all comments are SPAM, and that so far it has caught 410,599,317 SPAM comments, with 2,000,953 so far today… Those are impressive numbers. But this whole situation has discouraged me from writing comments — it is very frustrating when I write a (detailed and thoughtful) comment, for it to go into a black-hole.


[Note to the blogosphere: “Dear weblog owner/editor, unless you truly feel that I am SPAM, please feel free to re-classify me as otherwise” :-)

Cheers,
Carlos Enrique Ortiz