Striking balance for level 0 user

community
1

Hello, I help moderating a Discourse programming community at discourse.haskell.org. Today we got feedback from a user. I will summarise it:

  • He is a long time lurker.
  • He joined discourse to post an RFC for a library he was developing.
  • He could not post more than two links in his message (level 0 trust).

There have been some suggestions on the discussion:

But I am not looking for hard and fast rules, more about your opinion and experience on the inevitable level-0 vs. spammer friction and, if your community changed anything to the onboarding process, what you tried.

Thanks everyone and Discourse team for this very useful piece of software!

5 Likes
2

I’m hoping that eventually Discourse AI can automatically detect spammy / problematic posts from new users, so that most tl0 and first-day limits could be relaxed.

I do think the system could be presented more clearly to new users, but —

  1. Information overload at the beginning can be overwhelming and a barrier
  2. Spammers are insideous, and putting limits up front gives a road map for skirting those limitations.

Personally, when I recognize someone or just plain have a sense that a new user is acting in good faith, I manually bump accounts to tl1. That doesn’t scale, of course, but… it’s something!

5 Likes
3

i think a large factor in this is the type of community the particular forum is hosting. for mine, i definitely need those limitations on TL0 and even TL1. i actually use those lower trust levels as manual settings for when certain users request / need to have their access limited. perhaps more granularity is required in the TL settings? :thinking:

5 Likes
4

You could simply customise the error message to provide more info and a link to the TL doc.

7 Likes
5

i am actually unclear on what this refers to? why do you think there is any inevitable friction there or are you speaking only of your own forum?

2 Likes
6

My understanding is he refers to the unjustified and simply annoying restrictions “good” users have to go through (=“level-0”) vs. the usefulness of this system against “less good” actors (=“spammer”)

4 Likes
7

Indeed that is correct, the balance of “onboarding good users” vs. “keeping malicious one at bay”.

2 Likes
8

True, but those good users who are truely interested in joining the community will come back and overcome such resistance. Heavy filtering costs you bad actors(protecting your existing community) and some good contributors.

IMO the trust system is pretty good. Most of our users bump to trust level 1 after reading 10 topics and 15 minutes read time.

4 Likes
9

Have you experienced issues due to TL0 spammers?

My experience is that the auto-detection from Discourse, as well as Akismet, are almost flawless. I have basically 0 false positives.
Those who still post are quickly flagged, and their messages are hidden by the community, so it’s quite a non-issue.

Might be otherwise in other communities, of course.

3 Likes
10

Would like to thank everyone for their feedback! I will bring all of your suggestions to our mod team and see what fits better our community.

2 Likes