We finally did it! We built our own spam filter plug-in since we were having so many challenges with the existing spam filters, and we released it just last week! Being the geek that I am, this new plug-in has made me absolutely giddy! In my true, Boston-like way, I announced it to our internal team and told them I was "wicked excited", to which my boss replied, "You're such a nerd"! :-) I can't help it! The ability this plug-in gives us is far more intuitive, will give us the opportunity to identify patterns and adjust the filters, as needed. What a concept! With that being said, I want to break down what we've done so you are aware of the new requirements.
- Users must have (#) piece(s) of published content to avoid moderation
- One of the first things we did was set a limit to the number of posts a new user can create before they are able to bypass moderation. If you're a brand new user, your first three posts will be sent to our moderation queue for us to review. It's not that we don't trust you, but it will help us to ensure your question, document, discussion, etc. is placed in the correct forum, as well as make sure you're a real person! :-) This will also help us with false positives that we were receiving from the other spam filters...these false positives were why so many of you were getting moderated on practically EVERYTHING you posted! We don't want you to have to experience that again!
- Users with accounts (#) hours old will be moderated
- Another thing we have added was a timeframe associated with the account. This last spam attack we saw something different. We had users that were creating accounts days, sometimes weeks in advance, posting one thing that seemed ok and then much later nailed us with a script to post all that spam. By having this feature, I can say "Any user that is newer than 30 days will need to be moderated". I have this disabled at the moment because I intend to use this only when we really need it. I think having the 3 pieces of content moderated is enough for everyday filters.
- Flag users based on patterns using Java regular expression syntax
- Finally the best part to this new plug-in (and my personal favorite) is the use of Java regular expressions. With each attack, there was always some kind of a pattern, whether it be in their subject, email addresses, names, etc. Once the pattern is identified, we will have the ability to add a Regular Expression to flag those accounts so they can't even be created; they will go into a "Flagged User" group where we can bulk delete them once and for all before they even have a chance to send out content! WOOOHOOOO!!!!
Now that this new plug-in has been running for about a week, there are a few small changes that we want to make that will make our jobs a little easier, so we are going to take the site down for about 1 1/2 hours tonight. Don't be alarmed if you see a maintenance page tonight between the hours of 10pm CDT and 11:30pm CDT; it's all good stuff!
We hope that you guys are enjoying the new spam filters as much as we are, but if there are improvements you think we should incorporate or something we've forgotten, please don't hesitate to let us know! We love feedback!
*Side note: There are a few random posts that are holding on to some of the previous spam filters, so if you're moderated and it doesn't seem to make sense, don't worry! It will soon fade away and shouldn't be moderated anymore in the future (unless you post something on our banned keyword list ).