So, Lemmy is sometime missing content. I don’t regret switching from Reddit to Lemmy but, expecially for niche communities, the content isn’t always here.
My idea is to fix this is a Fediverse-based content relay named Relly.
Relly allows you to select RSS feeds, Mastodon users, Mastodon hashtag and Mastodon instances (so, the top posts on that instance) as sources for content, and post them to your favourite Lemmy community.
There are several features which make Relly better and anti-spam:
- Limits for a source (example: only up to 5 posts a day from this RSS feed)
- Limits for a community (example: only up to 5 posts a day to !archlinux)
- Global limits (example: only up to 10 posts made each day)
- Opt-out for servers & communities (instance and community moderators will be able to ask to be put in the UNLIST, which blocks by default Relly on your instace/community; this isn’t an anti-spam, as it is more a tool for avoiding common users to use Relly in a malicous and spammy way)
- Order posts (so, if i have 10 RSS posts and 10 Mastodon posts and a global limit of 15 posts, you can either have the 10 RSS posts and the 5 most upvoted Mastodon posts, or some RSS posts and some Mastodon posts [always the most upvoted])
- Multiple communities (post the same content to different communieties, or set up a fraction [ex. 50%], so that each post has a certain percentage to be posted on a certain community)
- Dynamic limits: You can set an objective of active users/post made in the last 24 hours, so that the limits (either for a specific source, a specific community or globally) will be reduced. Example: if you set a objective of 50 posts, and 25 are made, the limits of Relly will be 50% of what they were originaly set to be; this allows Relly to completly stop posting on a community if the objective was already reached.
- Do not repeat: before posting a link, checks if it was already posted in the community in a specific time period (by default, 48 hours)
- Modularity: new post sources and post outputs can be implemented; an example could be an e-mail output, so that you can run Relly in local and recieve an e-mail everyday with your favourite news)
Relly is designed to be used by moderators of communities, but users can also use it. A user should always ask the moderator if it is OK to use it. A moderator should always ask the admins if it is OK to use it. Moderators, if they are the one using it, should also make public the list of sources, and allow the community to discuss possible edits to the list. The admins should put in the sidebar notes if Relly is OK to use for moderators of communities.
At the moment, Relly is just the idea that I presented here; I want to hear the community’s feedback, and if the community is OK with this project being made, I will start working on it (I will make it in Rust and release under the MIT License).
The bots are a big step backwards imo. They make this place look like a surreal ghost town, instead of an inviting place to start or participate in a discussion.
It depends on the implementation. If all you have is a handful of bots posting all their posts on a fixed schedule, then yes, it gets quite tedious.
But I believe IMHO that the system I have going for fediverser avoids a lot of those issues. There is one “mirror bot” for each account that is posting on reddit, and they posted to lemmy’s mirror instance in near real-time. This means that the conversations basically happen in the same way they happened on reddit. This also opens up the possibility of (a) two-way communication between Lemmy and Reddit threads and (b) the mechanism for the real person to “take over” the bot account.
So what if people don’t want their answers being mirrored back to Reddit because it’s a greedy company, or redditors don’t want their comments on Lemmy?! Are you going to ask each one individually or just do it without their consent? Does it become a one-way mirror where someone from Reddit will never see your answer, or are you going to flood Reddit with bot accounts as well? That sounds really engaging and not at all creepy to me! /s
Good question. The way that I am designing it at the moment will actually ask for consent. The idea is that if you reply to a reddit mirrored comment, you get a bot telling you “hey, this poster is on reddit, connect your reddit account to if you want to bridge the conversation”.
Then it works by opt-out. By logging to the fediverser instance, the reddit -> lemmy mirror is automatically disabled.
So for someone who doesn’t want to use Reddit (probably quite a lot on Lemmy) these posts are gonna be filled with comments that they can’t really anwers to. That’s exactly what I mean by surreal ghost town.
I’m pretty sure that’s illegal in many places. You can’t just copy someone’s content and tell them to login to your service if they don’t like it.
There is an asymmetry here.
Judging by MAU, we have currently ~35k people on Lemmy because they are against Reddit. We have 300 million people on reddit who are on reddit because “it is where everyone else is”.
If the mirrors are creating 250k new mirror accounts per day, and if one fediverser instance can convert 0.1% of these per day, it’s 250 new users who “don’t care”. In one week, these ~1500 converted users will be in conversation on both networks, which will increase the number of non-bots in mirrored threads and be enough to stop the “ghost town” feeling.
I’m not talking hypotheticals. It is happening already some of the communities where I set this up.
I’m pretty sure you are wrong. The archive project does not need consent from the users (or reddit) to archive their content. Mirror sites from twitter exist for years already, none of them have faced charges. The worst that Reddit can do is to revoke the keys by claiming violation of the terms of service.