• Replies 12 replies
  • Subscribers 43 subscribers
  • Views 22341 views
  • Users 0 members are here
Options
Related

How to handle spam protecton in community areas

AngelaIp

Preventing spam is an important but also challenging task in maintaining quality in an online community.
Spam can be frustrating for community administratos, but also for the average community user.
You might already use some basic spam protection services, like a recaptcha or similar. But it's also important to check the content of your users' posts.
Of course you honor freedom of speech and censorship conflicts with your personal attitude. So what options do you have?

Your system may provide you a workflow that helps to prevent inappropriate content from being shown in the community. 
If you notice, that certain users repeatedly submit inappropriate content, abuse detection by counting forbidden keywords would maybe a good place to start.
Some systems may call this feature "Forbidden Word Counts".
There is a possibility that your system may provide this and other helpful features in Administration > Moderation > Abuse Automation.

Parents Reply Children
  • in reply to Suhas

    The problem is, that they high-likely already tried all my smartass answers themselves long before, but spammers constantly improve themselves just too fast.

    But of course that´s no reason to stop now:

    The spam content itself changed a couple of times already. Filtering according to forbidden word count don´t work for the current type of spam, which is mainly based on non-latin characters and doesn´t contain obvious forbidden words. 

    If there would be an general regex filter for non-latin characters, they would have tried already.
    I am not sure if the forbidden word-counts filter support non-latin characters. Worth trying would be:
    微信 --> would filter for 'WeChat'?
    or
    文凭 --> would filter 'diploma'?
    or
    大 -> just because it´s so common

    IP ban probably don´t work, cause it would affect a lot of innocent regular users. Disallow registration of noisy user names probably only would have temporary effect e.g. (akfgi14939).

    Same for reducing post frequency. What´s the right barrier? 1 minute, 5 minutes? It´s hard to use without restricting regular users too much.