[Feature request] More tools to fight spams

[mratsim]

We recently had a wave of elaborate spammers on both reddit and the forum.

Their question elicit good answers that are unfortunately lost to the community and Google after those spammer edit the links.

Examples:

• https://forum.nim-lang.org/t/5676
• https://forum.nim-lang.org/t/5675

I have 3 ideas to prevent that:

• Allow admins to lock a message for edition
• Lock messages for editions after a day as long as the original poster doesn't meet a certain bar (like 5 posts which ought to be enough to detect genuineness)
• Don't allow links or editing links in besides a whitelist (playground, github, gitlab, bitbucket, ix.io, pastebins, ...) if someone has less than 2 posts.

The first one should probably be there while the 2 and 3 are probably either one or the other.

[liquidev]

The two last rules could easily be mitigated though. A mechanism for browsing and reverting edits would be more useful, admins could browse through the older versions and revert to one that's legit a good question/post.

[mratsim]

Spammers usually work with scripts to target millions of websites. Adding special rules is probably not worth it.

Also putting a post threshold means more opportunities for them to get exposed. It is easy to pass yourself for genuine with 1 post, just ask the same question as everyone. But 3~5 genuine posts is much harder and for admins that makes more post history to base future decisions on.

[mratsim]

Other antispam tools:

• currently we use reCaptcha but there are Python scripts available in the wild to combat that.
• We can add:
  • an extra field with a simple math equation
  • something that is not math "write 20 without 0"
  • a hidden honeypot field that if filled, drops the connection. (Note some care should be given for accessibility but are reCaptcha even considering user accessibility?)