What would be a good way to deal with duplicate posts?


(Drew) #1

On the Discourse forum I use, we see a lot of duplicate posts, and it really is showing on thread about a new feature.

There are currently 54 posts on the thread. A staff member on post #17 listed all of the known issues regarding the feature. If the topic is summarized, it’s only 11 posts and the summary includes #17. After post #17, we’ve averaged 1 every 6 posts where someone has complained about something listed under “known issues”. If we don’t include the “That was already mentioned in post #17” replies, it’s 1 in every 4 posts. The staff members could (and should) edit post #17 into the original post, but that’s not the problem we’re looking at here. The problem is that people aren’t taking a couple minutes to read the thread before posting, and that results in both a lot of duplicate posts that make the thread hard to follow and circular discussion.

I posted a topic on the forum as sort of a call to action for moderators to start deleting these duplicate posts (maybe by editing them to be [Duplicate Post] so the poster knows why it was deleted and then leaving the poster to delete it themselves), but other users of the forum said “You shouldn’t try to program people’s behavior” and “People would stop posting because they’d be afraid to risk being moderated”. I don’t see the deletion of duplicate posts as any different from deleting personal attack posts (in both cases you’re just deleting destructive posts), so I’m not sure I follow on the programming behavior point. Regarding the scaring people off because they fear being moderated, my outlook is that they’d have nothing to fear if they read the thread before posting and good riddance to any posts lost due to people not wanting to read the thread because their lack of effort would have likely reflected in their posts.

It’d be one thing if people didn’t read a 200 post thread, but the topic in question isn’t that long and the Discourse software estimates it’d only take four minutes to read the long version, and it’d be significantly shorter if someone read the summarized version, so there’s no excuse for people to not read this particular thread in its entirety before posting.


I’m wondering: are the fears of behavior programming and posting scare off without merit, or are they actually an accurate estimation of what would happen if duplicate posts started being deleted? If those concerns are real, what would be a better way to deal with duplicate posts?


(Jeff Atwood) #2

This is just weird – why summarize the issues way down practically in the middle of the topic at post #17? Who’s going to scroll down there, or press the summarize button… and even if you pressed the summarize button, you’d likely still need to scroll down.

First step: clearly summarize the issues in the first or second post. Right at the top where new visitors to the topic can easily find it.

Anything else we could discuss… honestly, until you’ve done that, there’s no point.


(Drew) #3

The staff members have since added post #17 to the original post. The reason I didn’t pay much attention to its location was because it doesn’t have to be a known issue posted by a staff member. It could be another user posting a concern at post #17 and then other users repeatedly posting duplicates of post #17 instead of liking it because they couldn’t be bothered to read the thread. The problem we’re looking at here is people not reading the thread – not the initial choice of location for the “known issues” list.


(Jeff Atwood) #4

In that case, delete the duplicate replies and post a staff message that tells people to be sure to read the first post – and make it very clear in the first post that new replies which clearly have not read the first post, and post duplicate information, will be deleted.

Deleting the dupe replies also keeps the topic shorter, so more people can process what’s in it. It’s a good idea.

We’ve discussed speculative future features like requiring users to have read (scrolled past, or have cumulative seconds of read time) before they are allowed to reply, or placing the number of replies they’ve read and cumulative read time next to their reply, ala Steam Community…

… but that’s fairly draconian and we are hesitant to put anything so severe in.


(Drew) #5

What of other threads? As in,topics not created by staff members. The recent staff-posted feature discussion topic was just a good example of a problem that mostly happens in user-made threads. It’d be pretty high and mighty of a regular user to post “Guys if you don’t read the thread and make duplicate posts / start circular discussion your posts will be deleted.” If this were to be applied to the forum as a whole, it’d have to be a general rule that everyone knew by heart. We arrive back at the two concerns users of the forum had with that: deleting duplicate posts on topics throughout the forum is trying to program peoples’ behavior and would scare off discussion. In your experience, do those have merit, or are they empty concerns? If those are real problems, then we obviously couldn’t apply the post deletion on a large scale.

Maybe add a “duplicate” auto-flag? If too many posts (% so it’s not biased to large threads) on a topic are flagged as a duplicate and deleted, there’s clearly a problem with people posting before reading. Once it’s become apparent a thread is habitually not read, maybe then and only then activate the require-you-to-read-before-posting feature? Not sure if that’d be too stringent or not on such a specific application basis.


(⛰️) #6

All of them have merit. Priorities must be set clear on what is most desirable; mitigating (or outright extinguishing) the utmost undesirable of behaviors, fostering others, through any means those in power wish to do at their Will; to come about the conclusion as to how they, and you, see this community.

This is an answer, I feel, that leans toward structured, groomed content.

Depends on the community in any given context, if that is more valued. Deleting content for the sake of keeping the presentation orderly can swiftly veer into–

–that kind of territory. This I have seen. The most pathological form comes from crowd/social psychology as the spiral of silence.

I only offer this viewpoint for future readers to consider the “collective consciousness” of their community overall, while also attempting some form of order to the topics themselves.

The default FAQ has a good start to this. Duplicate content is not specified (“Keep It Tidy”). Flags can be used too.

Lots of options to consider from tools already in place.

Then again, I do not know the specific context here for the community mentioned; the social mentality expected, upheld by the staff and/or the users themselves.

Lots of variables, I see.


Youtube scrolling bug