Use moderated tags instead of subprismos


I think Prismo should do away with subprismos and use moderated tags instead, where admins can set special rules for a particular tag (e.g. #avengers and #nospoilers) and assign that tag to a mod team.

My main reasons:

  1. There is a lot of overlap between the two concepts. I think having just one way to categorize posts thematically will make things simpler in the long run.
  2. With subprismos you have to choose which community you want to talk to, and that splits the conversation.
  3. The existence of subprismos can create indecision; if a user wants to create a post discussing a videogame-based movie, should she use the !videogames sub or the !movies one?
  4. The existence of subprismos can cause people to create subs mimicking popular subreddits, even if they aren’t really interested in the topic, just to become mods and have power over a popular sub.

Here’s a couple mockups:

When a user creates a story, she can give it as many tags as she wants. Moderated tags appear colored and there’s an indication that there are some additional guidelines, and that she should follow those or remove the tag.

If she doesn’t follow the tag guidelines but still uses the tag, other users can report the story. Tag mods can remove the tag, and they can ban that user from using it in case of abuse.

If the story respects the tag guidelines, other users’ comments also have to follow those guidelines; if a comment doesn’t follow them, it can be reported and have the tag removed, which hides that comment (and the whole branch it originates) from view.

Then, users who open the story see a list of the moderated tags, and a disclaimer that some comments
were hidden because they don’t meet their guidelines. Users can then remove specific tags in order to show those hidden comments, or also if they want to write a comment that will go against those tags.

That way a user can visit a #nospoilers story and write a spoiler, but in a way that users won’t see it unless they don’t mind spoilers, and mods won’t have any extra work.

I think this approach has a number of advantages:

  1. Navigation is simpler, because there is only one way to categorize posts by topic. Not two.
  2. It makes more people join the conversation, because it’s not scattered across separate silos.
  3. It makes it easier to create a story, because the user doesn’t need to decide where to do it. If he wants to create a post about a movie based on a videogame, he doesn’t need to choose one sub over the other; he can simply choose both tags.
  4. Related to number 3, it makes it easier to find the stuff you want to find. If you want to keep up to date about your favorite show, but don’t want it spoiled, you can simply find where #gameofthrones and #nospoilers intersect.


One issue that was mentioned in the matrix room was that we might be alienating subreddit users by not having subprismos. The idea being that having only tags might be hard to explain to new users who are so used to subreddits.

I disagree with this idea. I think there are essentially 3 options:

  • Tags only;
  • Subprismos only;
  • Tags and subprismos;

Between the first 2 options (only subs or only tags), I think having tags only is, from a user standpoint, the simpler option. The usage of tags is much more frequent in websites than the usage of subs. Even instagram has tags. Between one of the first 2 and both (tags and subprismos), I think having only one of them is the simpler form, for obvious reasons.

And, as far as the user is concerned, tags behave pretty much the same as subreddits. If you want to view posts for a specific tag, you go to


that’s true but when i’m thinking about a regular internet user, he/she probably can’t imagine how tag can be private/restricted/moderated (but maybe it’s simply me being too protective). For such scenario, they are more used to facebook groups


I think that would be easy to solve. When a user chooses e.g. #science, she gets a little disclaimer underneath the tag list. For example:

“In order to tag this post as #science you have to follow the following rules. Please, take them into account if you want to use the tag.”


At that point, wouldn’t the only practical difference be that posts can have multiple tags?


No, there is another practical difference that is, in my opinion, severely annoying and makes me not really like this tag idea afterall. And that difference is that with subs, you have several vote counts. If you have 2 posts in 2 different subreddits, you woulld have 2 different vote-counts. One for each sub. That would not be true for tags.

Imagine a post tagged with #pics and #science. The post gets massively upvoted by the #pics folks but not upvoted at all by #science readers. Still, that post would be in the frontpage of the #science tag.


and there is also another issue, the one of blocking/banning tags by a user. You can do that easily with subreddits, just unsubscribe from it. But would that work with tags?

Imagine, again, a post with #science and #pics. You hate seeing pics and do not like the comments of the general viewer, so you block/ignore/unsubscribe from #pics. Would you still see the post that was tagged with #science and #pics?


TL;DR : I think we are missing the point that Prismo is federated, and that changes things. I believe in tags, because they are an existing concepts, and Prismo could be a tag-oriented prismo on the fediverse. This is in my opinion the most compatible way to do it, and it ressembles more a federated HN with modern tools for federated moderations/administration than a federated reddit.

Generally speaking, I think the discussion is currently missing one huge point: Prismo is meant to be federated, and largely distributed. This means that anything that is to be tightly controlled must be located, i.e. bound to a specific instance.

This is supposedly what sub-prismos would bring: a way to create a local sub-version of the prismo site, with its own rules, own theme. Of course, sub-prismos would not be a rare resource, because anyone can open an instance and have any name available there. That does not mean that a race for power would not happen just like on any other reddit clone.

But this is still hugely missing the point: Prismo is to be federated and distributed. So why create sub-prismos if they are meant to be small specific communities, with their own rules and own theme? Why not simply have more Prismo instances, that can have different rules, and different themes, and different mods, and different whatever?

Now comes the point for tags. Tags are an already-existing concept in the fediverse and ActivityPub, which has the pretty nice advantage of already being compatible with pretty much any other implementation, including Mastodon, Pixelfed, etc. Of course, tags have a huge disadvantage: there is no tag moderator, there is no single power in charge of maintaining a tag because it is not bound to an instance ; thus, local moderators simply enforce their local rules on each instance. But hey! isn’t that exactly what we are trying to avoid, a race for power on rare concepts that are sub-“reddits”? If you do not like the way you instance admins or moderators are removing some posts because they are mistagged, simply create your own instance, no central power, no race, more subtle and more flexible rules.

Then, one must argue that it will be very difficult to enforce tag meaning in such a world. For instance, how do you make sure that every #science post is indeed about science? I believe you can accomplish this using local, per-instance moderation. Currently, tags are really second-class citizens on the fediverse ; Prismo could make them a first-class concept because Prismo could basically be the fediverse from a link-aggregation and tag perspective. Once tags become first-class citizens, people, and moderators, will hopefully care more about them. And people will try to define collectively standard meanings for tags. Some tags will be informative, like #science and #sport, where effectively no specific rule can reasonably be enforced ; some will probably be more constrained, like #academic, and people could agree that only posts citing their sources should use it. Some tags could even have positive rules, like you should tag #nudity a content that contains nudity. This would help solve the huge problem that are currently content warning on the fediverse, and that barely help setting up safe instances, let’s say, for children.

What would then happen to moderation? We would need more tools, and more subtle concepts indeed. For instance, we would need to be able to mute or block a user, or even an instance, for a given set of tags; you could then decide that people from that Prismo server (or Mastodon, or whatever for that matters), have abused the #academic tag too much by not citing their sources, and any posts from them using these tags should ignored for a week. Of course, another Prismo server might not agree with you, but that’s only faire because moderation is local to your instance.

Also, what the fediverse is currently missing is a system of trust and spam distributed database. A way for an instance to publish the list of people they do not trust, at all, for a specific tag, or for any other reason. And a way for other instances to simply subscribe to these, when they do not have the time or will to do moderation themselves but do trust others to do it correctly.

Finally, I see the concept of private sub was mentioned. It seems to me that a private sub is simply a limited audience for a post. I remember there was some discussion recently in the Mastodon dev team about providing tools for targeting specific audiences, just like GS groups. Maybe this could be a great way to be headed, in order for Prismo to be compatible with the whole fediverse.


I think a hierarchy is great from a UI/UX perspective.
(Instance on Machine Learning / Subs on eg Bias)

Not everyone will want to create a new instance, but many will want to group posts and discussion in one place.

Subs are somewhat moderated/curated, tags will never be. Tags are for global ref, Eg when searching for subs about bias in machine learning it will be useful to be able to find #ml #machine-learning #bias, and nice to be able to then follow Subs that you locate on the topic.


Bottom line:
Go for tags and subs.
Available UX elements on every instance.


I suppose in that example the story meets the rules of both tags (e.g. it is a pic about science) but the #pics crowd votes it because it’s funny, while the #science crowd usually vote stuff that’s academically interesting.

I don’t see it as a huge problem (after all the story is about science) but some people might want to avoid it, and tags would offer several ways to do that.

The most obvious way would be to create a different tag (e.g. #academia) where pics don’t belong, and follow that tag instead. If the number of people who want to avoid your scenario is big enough, they will soon start tagging their stories #science and #academia.

Another way is to ban the #pics tag, in order to only get non-pic stories in your #science feed. Or if you don’t want pics in your #science feed, but you want them in your #cats feed, ban the intersection #pics/science.

I understand this is a very specific solution, but I believe it’s a fringe scenario too.

No. I think that if you hate seeing pics, and ban #pics, you shouldn’t get any more #pics stories in your feeds, even if they’re about science.

Then again, if you think a user might not want to see cat pics, but he might want to see science pics, we could let him ban the intersection #cats/pics to only stop seeing that specific kind of pic.

But that’s exactly the point: to make tags moderated :slight_smile:


I was thinking about how this whole idea of moderated tags could work with federation, and I think I found a way. Feedback is welcome.

Imagine three Prismo instances, each with a #science moderated tag. The people moderating #science on Instance A and instance B agree on some common rules (e.g. no flat-earth stuff) and decide to federate those rules, but the mods on instance C decide that they will allow flat-earth stories.

As a result, if a user from Instance A reads a #science story from Instance B and decides to boost it, the #science tag is automatically applied, because Instance B is trusted.

But if he reads a #science story from instance C, and decides to boost it, the #science tag is suggested but he’s asked to check that the story meets the #science guidelines on Instance A.


I propose implementing a web of trust for moderating tags for thw fediverse as a whole. That way moderating tags can be done in a decentralised way that also commondates moderation of the fediverse through collaboration and allows any end user to be in control of their own views of how/things can be moderated. The downside is.segregation of trust networks, but that problem already exists in the fediverse in a sense and is a complex discussion in itself.


I think that’s missing the point.
I want to follow #science tag/subprismo, and I don’t mind if they are picture posts or not. So, I don’t mind a post that belongs to both #pics and #science, but if it hasn’t gotten any votes in #science (or negative votes), I don’t want to see it on top, I don’t care if it has gotten a few thousand upvotes in #pics.


If I understand, you don’t mind seeing it in the #science feed but you don’t want it at the top of the feed, because while it technically belongs in #science, it’s not the sort of post that’s popular there.

Is that correct?


I think there’s room for both. Single-sub instances that mimic HN and multi-sub instances that mimic Reddit.

Within each sub you can still have tags, but it is the sub that is moderated, not the tags. Subs/instances with a broad scope would use a wide variety of tags to categorize content, while narrowly focussed subs/instances would probably use fewer.

The barrier to entry for creating a new instance might be low from a technical perspective, but it still goes over the head of the average end user. There need to be places where I can create a new sub on an existing instance so that end users have that option.

Tl;dr is that I agree with @arjen. Instances have one or more subs, and tags apply within the scope of a sub.


Yes. For bigger subreddits, most users generally don’t browse all posts, they just go through top few (10, 20 or whatever) posts depending on time / interest.


So their experience would be the same if the pic story just wasn’t in the feed, as opposed to being down below?


Yes, but they can view it if they go down enough, it all depends on how much votes it got compared to other posts.

The main point is, how do you differ between #pic votes and #science votes, because there can be another post, with both #pic and #science tags, which got high votes in #science so I want to view it.

I don’t want to filter out any posts, I just want to sort them based on it’s popularity in #science.


It was discussed in the matrix room that maybe the vote-count could be normalized to the popularity of the tags for a given period.