Should Discourse require reading to gain initial trust?


(F. Randall Farmer) #1

I just went through signing up and trying to post something with links/images at ENWorld. They have a “must make 10 posts before you can post images/urls” rule. That was very painful and has lead to people running around posting low-quality posts to arbitrary threads.

The worst part was that the system gave you no feedback along the way. It didn’t say how many posts you needed to make. The UI didn’t block-off unavailable features (you only learned about this when you attempted to submit) - this was very frustrating. The “URL” or “Image” button could have thrown up “You can’t do that until you make X more posts.” This on top of their 90-second rule for searches was a royal pain to find something to contribute to to meet the bar. All to avoid spammers.

The idea of an optional peer-approval process for n00b posts (and allowing URLs/Links in posts that requires this approval) is growing on me. Keeps the spammers at bay (ala enworld) but allows known/demonstrably-good new contributors to quickly contribute great content and advance their trust rating.


What do user trust levels do?
(Jeff Atwood) #2

With all due respect, I think these are very different situations. Compare:

  • you must create 10 new posts
  • you must scroll and read… stuff

The former is many, many orders of magnitude more onerous than the latter. Reading is completely easy and passive. Just scroll! Reading is fundamental on this and every other forum, just fall down the gravity well with us.

Forcing people to create posts is, pardon my French, really stupid. It encourages talking without listening, the worst kind of “post count” inflation nonsense, and posting is not even the core activity on a forum – reading is!

I am not sure it makes sense to show a countdown timer like on an E-Z bake oven, but I am certainly for showing more just in time feedback when you attempt to do something that you are not allowed to do based on being a new trust level 0 (visitor) user.

Anyway, it’s rare that a new user, who is just getting their footing and exploring the community, and isn’t a spammer, would even need to do things like post more than one link, or an image… remember that the new user levels are there for the new user’s protection almost as much as the forum’s protection.

I don’t mind the idea of approving new user content manually, but I also think what you describe (you must post 10 things!) is a terrible, apocalyptic scenario, that can only end in tears. It would have me clawing for the walls in escape too… just awful.


(Jeff Atwood) #4

Requiring 10 posts is not the same as asking users to introduce themselves, though!

And if there were introductions I would want them walled off in a dedicated topic for that…


(Guillermo Borges) #5

Or category.

This post size limit is getting silly…


(Guillermo Borges) #6

To explain myself further:

I think having a system where people can be upvoted into trusted status would be great, and it would encourage people to make more elaborate introduction topics. It’s also very hard for a spambot to get through.

Not sure if this is already a feature =o


(F. Randall Farmer) #7

Agreed completely. Just had a little exchange with Morrus on the matter. The problem is that this requirement is stopping the spammers for him at the moment - so he’s moved on to other fires.

Hope no one thinks I was saying the 10-post requirement was good in any way! Yuck!

Do we allow one link and one photo at level 0?


(Alex R) #8

Up to 2 links but no photos, IIRC.


(Lee Dohm) #9

From What do user trust levels do?


(Vitor Eiji Justus Sakaguti) #10

I am currently a level 0 here at meta.discourse.org and I have no idea what I still have to do to be a level 1. There’s no hint at all in my profile page. “How shall I prove my worth?” I ask myself, but the answer is unclear.


(Jeff Atwood) #11

We’ll improve that soon, but the answer is “keep reading!” for now.

Reading is Fundamental!


(Vitor Eiji Justus Sakaguti) #12

Good to know!

but the answer is “keep reading!” for now.

Reading is Fundamental!

Yes, I got that :slight_smile:
My real question was “how much reading”, but that will come soon, as you say.

By the way, it seems I’m a basic user already. Nice!


(chronodekar) #13

I agree with @codinghorror on the usefulness of reading to “advance” a user. But here’s a scenario that I find myself in,

Browsing the internet on my work PC. I spend a good amount of time just reading stuff and on the occasion that I want to post something, I end up going home, making an account and then talking. The problem now is that I’ve lost whatever “reading” time I spent earlier.

Granted, it’s probably a niche use-case that could probably be nullified if I just log-in the next time I browse the site/forum from work (or another system), but its still a nit-pick.


(David Miller) #14

I was thinking roughly the same thing: typically when I’m researching a topic for the first time I will come across forums in search results which I will read before wanting to post. (As codinghorror has said, it’s generally better etiquette to read existing forum content before posting.) I don’t like creating accounts unnecessarily, and most forums* only require users to be logged in for posting (not for reading). If I did this on a Discourse forum, I would lose my ‘reading credit’.

With your single sign-in support, creating an account of course becomes far less onerous - but it’s still not necessarily something I’d want to do immediately on first visiting a forum.

Given the client-side cleverness of Discourse, would it be possible to track ‘user reading’ of a forum before the user created an account and apply its ‘credits’ to the account on its creation? This wouldn’t solve chronodekar’s problem, but it would solve mine.

*I have come across forums which hide links or images for users that aren’t logged in. Personally, I find this intensely irritating and tend to look for other forums!


(David Miller) #15

On a related topic, perhaps, when I tried to post my reply above I received “Body has too many links” until I removed all of my @ references to chronodekar and codinghorror. This is presumably because I’m not yet sufficiently trusted! Perhaps I was wrong to do this, but it seemed natural to refer to people using @ links?


(David Smith) #16

An aside, but I find the phrase “entering at least 5 topics” ambiguous. I think you mean “viewing” but it could be interpreted to mean “starting” (think of “enter your name”).


(Sam Saffron) #17

Well this is tricky business.

We don’t want brand new users @mentioning every user in the forum in one post, there needs to be a tighter limit for visitors (maybe allow them to mention up to 2 users), but we could have separate limits here… eg. 2 links AND 2 mentions … or something like that.


(Jeff Atwood) #18

There is already a rate limit on maximum @ name mentions, but nothing specific to visitors.

I think the bug might be that @ name mentions are counted as links for visitors? Something else we need to check @eviltrout, that might explain some of the complaints if true.


(David Miller) #19

Ah, I see - an intentional limit so that new users don’t alert multiple users (as I nearly did). This makes sense. The @ syntax is presumably intended to direct a question or response explicitly at a user when it might otherwise be missed…


(Robin Ward) #20

There was definitely a bug with @ name mentions being flagged as links. It’s been fixed. I’ve also given much clearer errors when you hit the maximums of links / images / mentions. It tells you how many you are allowed. And we increased the max link count for visitors to 2.


(Ides) #21

Similar to user names, couldn’t discourse contact the main discourse server to help verify the user to be able to do a little bit more then what a user could do if they created a completely fresh account?

This wouldn’t apply to discourse if they disabled the ability to contact the mothership or decided that it wasn’t secure enough but it would solve the issue for the average discourse you come across.

I’m not exactly sure how the mothership will be able to make a very reliable way to certify a user is not malicious unless they are active on meta forums or other fairly official discourses.