Vanilla 1 is no longer supported or maintained. If you need a copy, you can get it here.
HackerOne users: Testing against this community violates our program's Terms of Service and will result in your bounty being denied.
Spam and vanilla
Hi guys,
This is officially my first post here, and I wanted to have a good start. So let's talk about spam, baby. :)
Another (phpBB) forum I moderate is being flooded by spam, especially these last few weeks. SInce I plan to install Vanilla for a new project, I was wondering how spam proof Vanilla actually is. My research leads me to believe spam is a bit of a non-issue here?
1. There are only two anti-spam extensions:
http://lussumo.com/addons/index.php?PostBackAction=Search&Keywords=spam
2. And I could find just 30 threads with the keyword spam:
http://lussumo.com/community/search/?PostBackAction=Search&Keywords=spam&Type=Comments&btnSubmit=Search
What have your experiences with Vanilla spam (prevention) been?
0
This discussion has been closed.
Comments
With regard to malicious user-related spam - ie. the type of spam that happens when someone tries to post the same message a ton of times and flood the board with bs - Vanilla has a built in mechanism to place blocks on people's accounts if they post a certain number of comments within a certain number of seconds. It's completely configurable through the settings tab's "application settings" form.
I think manual approval is still one of the best ideas. However, it is just very nice to have a fully automated solution.
[brainstorm mode]
In the future, a combined automatic and manual approach would suit me best. Maybe something like Akismet filtering to filter out real spam combined with some Bayesian filtering to detect suspect postings, which would then be presented to an admin to manually resolve the spam/ham question? I know, that is somewhat of a fantasy. However, Akismet is available, and for Nucleus (a blogging tool), we have a plugin for Bayesian filtering of comments (link).
[/brainstorm mode]
There are NO PERFECT SOLUTIONS. Period. Yes, spammers CAN get past email verification. They CAN get past captchas. They CAN get past Akismet (and they do on regular occasion on my wordpress blog). Heck, Mark even points out how they can get past hand verification! I have not yet seen a single spam solution that is 100% effective because there is no such thing -- just like there is no such thing as perfect encryption.
However, each additional hurdle is one more thing that it takes them time to break -- and that is the deterrent. Why bother with your site when one IP address away some noob just setup an old phpbb install? Woot!
Now, I don't know what kind of super massively popular site you run, but I can tell you from running some fairly large sites (20,000+ users per site) that the number of bots who can complete email verification has been about -- well, none that I have ever seen, actually. Secondly, I could show you an encyclopedia's worth of log files that show spam bots being bounced by a simple Javascript automated captcha. So, I do agree that these things can be broken, but I never see them go the lengths to do so.
NOW, some solutions for you:
You say that you can't go with hand verification, I assume because you'd find it too tedious. I can see that. You probably don't want to use "invited by a friend" type feature either because that would lock out outsiders who didn't know anyone. Why not make a mashup of the two? Allow people to apply, allow all users to see the applicant list, and allow users to sponsor other users into the site. This way you use your current user base as a spam bot detector and can take away the privledge from people who prove to be very bad at it
If you really think Akismet would be the solution, put it in! It's very easy to make an extension that can hook into the comment/discussion pre & post posting actions. You could just jab akismet in right post part and have it scan the comment and hide it if it comes back sour. You'd want to be notified that this happened, of course, but that's not too much work either.