Does programming.dev have a dilution problem due to too many communities?
Does programming.dev have a dilution problem due to too many communities?
So, I love this site. I've been here more-or-less since the beginning, across various accounts. I also have accounts on other Lemmy instances.
One common pattern I see is that instances branch out their communities too soon, and overly dilute the conversation. It makes an instance that is ultimately not that active (compared to any of the big sites that don't need naming, really) appear to be even less lively, due to so many instances with either nothing at all, a few month old posts, or a generic post linking to a projects blog.
Note that I am not criticizing the instance by pointing out the low activity levels - I really do love this place. It's just a fact at the moment. You can switch viewing posts by new and scroll down a little to see we get around 5 - 6 posts per hour, occasionally a bit more and occasionally a bit less.
I think that having lots of inactive, dead looking communities is off-putting. I know that I certainly don't feel encouraged to post in them. I worry this might have a similar effect on other users too.
I do understand that c/programming is deemed as something of a catch-all community, and so anyone could post there rather than the niche communities, but I'm not sure that this is totally obvious to everyone.
Personally, I feel we should purge all the tiny communities that have no posts (or just a single blog post, for example) and encourage people to post in c/programming. Then, new communities can be made when a particular topic becomes large enough to warrant divergence, either because it's clearly a subject of interest to many users or because it ends up dominating c/programming. c/rust is an example of such a community, as is c/programmerhumor.
I am nobody here, and I was not asked for my opinion, but I just wonder if this topic has been thought about much? I really want this place to thrive. Do any other users here have an opinion? What do the instance admins think?
I think this is a pretty big problem with how people are using Lemmy in general
https://lemmyverse.net/communities shows 27,726 communities
https://join-lemmy.org/instances says Lemmy has 42.2k monthly active users (as of v0.19.0 that's including people who only vote and don't post/comment)
I just think that's way too many communities to be sustainable, people jumped on Lemmy and tried to create a community to match every subreddit, and then they did it on multiple different instances too.
I'm not sure how to improve this. You could delete communities but then you lose post history and subscribers. You could close communities and make a pinned post, but then they clutter search results and they definitely look dead, because they are. There could maybe be a new Lemmy feature added to merge a community into another one (moving the posts over and then deleting the old community) but that would have complications with subscribers on other instances.
We switched up how we handle communities a couple times. We didnt start expanding horizontally until the main communities seemed stable and had some good activity (community creation was open publicly starting last month and before that there was the request zone). The posting activity is the same as it was in the previous systems, people just dont want to post and theres the 10% of people comment, 1% of people are posters rule (if its 5-6 posts per hour thats better than the past system as well)
The ones that are visible in new communities are empty mostly because they were just made (~ 100 coms made in the last month). The ones made by me and that are now handled by the vacant account are ones that have sources I can add to my rss reader and then post consistently. For example the elm community has weekly posts on things going on in the ecosystem. This means they wont be dead and should have a solid stream of posts as it builds up active users. Theyre made so they can start collecting people and subs over time rather than attempt to diverge and have nobody know to move over or people not joining at all since they looked for the language by name and didnt find anything (and then I assume people dont want misc changelogs of bug fixes posted to c/programming as well)
The most commonly used sorts in the instance are subscribed sort and local sort. If something is posted in one of the smaller communities people will interact with it in local sort. Crossposting is also highly recommended. I can maybe somehow make that more apparent than it already is
I understand. Do you think that making 100 communities "to start collecting people" is a bit counterintuitive, though? Just a quick browse through shows that the majority of these coms (including ones made much longer ago) are just dead. A lot of them have moderators that haven't shown signs of activities in months, and the only posts are the RSS style feed dumps, with little sign of discussion. Would it not make more sense to just let people make communities who are interested in actually running the community? Including starting discussions, advertising the community, sharing interesting content within it, etc.
Please don't take that as a criticism. I see that you do a lot around here, and I really appreciate your efforts. I'm just concerned that this mass dilution may hinder a lot more than it helps.
Even looking at some of the larger coms, like [email protected], haven't had activity in over 10 days. In my opinion, that makes it feel unappealing as a place to go and discuss and share things about that topic.
The 5-6 posts per hour I mentioned is a little disingenuous, too. Looking through it, half of those are a single user (Mac), who is just going through these empty communities and posting links to the project's news feed.
Do you not think there would be some merit in having fewer, but more condensed and livelier communities?
I may well be wrong, I am certainly no expert in creating or running something like this. I am just drawing from the experience I've seen of places like old Reddit (back when it started) growing. They didn't let anyone make subs until around two years in, at which point they had reached a critical mass of users that meant fracturing into subreddits didn't leave the whole site feeling thin.
I think the better option rather than condense things into less communities is to crosspost things between the larger communities and the smaller communities and make the larger communities more apparent to funnel people into them
We tried the system with people interested in running it with the request system, it didnt work and people didnt actually boost things
Mac is my posting account. I can start crossposting a bunch of stuff I post on it into the general communities if the specific things have less than 100 active users and do some more discussion posts
I don't disagree, I do think there are too many communities for the number of active users (both here, and Lemmy in general). What I'd be interested to know is this: Is there some research into the subject, or even a write-up from someone who has successfully grown a thriving community in the past?
I'd argue that with [email protected] being the "default" community, this is somewhat mitigated. It's not policed, so you can post there about Rust, Godot, Python, or whatever you like and nobody will moderate you or ask you to move along. Maybe the "over-dilution", as you call it, hurts the instance as a whole. But if you think of Lemmy as something wider than a single instance, it matters less. [email protected] is the flagship instance here, and it's a large one by Lemmy standards. People will subscribe to that from all over the Fediverse.
So I think it comes down to your view of programming.dev as an instance vs Lemmy as a network of federated communities. Ultimately, people will just subscribe to whatever instances interest them - and hopefully Lemmy as a whole will thrive, including this instance.