Email or username:

Password:

Forgot your password?
All posts Julian's posts Post Back to profile
Top-level
Julian Lam

We also have integrated project honeypot, akismet, and captcha, but those are passive technologies that can be worked around given time.

But simple is best — raise the bar of contribution just a tiny bit, and the effect on spam is huge. To get around a post queue, spammers have to make a quality contribution, and that's hard when spam accounts are automated by script.

Does #mastodon support any of this?

16 February at 14:56 | Open on crag.social
7 comments
Julian Lam

The silver lining to this: if you have this enabled by default and a server exists with poor or absent moderation... well, no spam gets approved 👋

16 February at 14:57 | Open on crag.social
Multiverse Mike

@devnull

I LOVE that you are doing your own thing, and love that you are not only questioning what has been "acceptable" up until now, but also creating an implementation to do exactly that. Keep up the good work! 🙌

16 February at 15:23 | Open on toot.wales
Julian Lam

@multiverseofbadness thank you!

The fact of the matter is, a lot of popular #fediverse software is rather new, and that's super exciting, but also means some of the software is rather green and not battle tested.

#nodebb went through its own growing pains, so I feel like there are some lessons that can be learned from outside orgs looking to join #ActivityPub devel.

I'm not saying we're perfect, but if we can help soften some of those growing pains for all of #fedi then 👍

16 February at 15:27 | Open on crag.social
Todd Alström

@devnull It not, it should. Puts a bit more work on the mod team, but stops nearly all spam. Using a similar option with XenForo.

16 February at 14:59 | Open on mastodon.social
Julian Lam

@toddalstrom all of this has been figured out by the forums 10+ years ago. The spam is just getting more sophisticated and larger in volume, but the old solutions still work because they introduce a second DIFFERENT element to the spam equation a human.

You can beat tech with more tech (e.g. CAPTCHA vs CAPTCHA) but you need to leverage social engineering against a human. Not infallible by any means, but a huge sharp turn away from what beats tech.

16 February at 15:40 | Open on crag.social
Todd Alström
Daniël Klabbers

@devnull @toddalstrom I can vouch for this. We have been finding brilliant ways to prevent spam in #flarum, including trying AI to filter. But early flagging is definitely one of the most effective ways to tackle spam.

16 February at 16:40 | Open on fosstodon.org
Go Up