When I started my personal instance, one of my goals was to run it cheaply and with little maintenance. Being writing Ruby for many years of my past life, the decision between Mastodon (Rails) and Pleroma (Elixir) was quite easy. A week later, my Pleroma instance is taking 500 megabytes of RAM and a few percent of the CPU of the second cheapest Hetzner instance (4.5€/month).
Even better would be a deployable binary service. Calling out all Go and Rust developers!
Huh- Running Postgres at home, while running the web server on Digital Ocean. I would think that since running several round trips to the database for every web request, it would make more sense to run the entire web server at home?
I suspect that a lot of folks are in the same boat. Mastodon will be getting a workout. The bugs will show up en masse, but they will also be fixed, and the service will improve greatly, as a result, with a lot of “lessons learned” lore, added to the canon.
I noticed that CNN ran a front-page story on Mastodon, so it’s now mainstream.
On a related thread (https://news.ycombinator.com/item?id=33562163) there was some discussion about the fork-centric model of Mastodon codebases and instances, so "will also be fixed" is one of those nuanced things: they may be fixed in https://github.com/mastodon/mastodon but how long will it take to fold those patches into the forks and update the hundreds(?) of instances out there advertising their version number on the homepage?
Twitter's biggest/hardest problem seems to be content moderation. With all the people moving to Mastodon now, it's not clear how Mastodon is going to handle this?
Is the hope that just folks of a given political viewpoint will move to Mastodon and moderation will not be necessary?
On a server level users can be reported by users and banned by admins. Mastodon is federated, and individual servers (or even individual users on servers) can choose not to federate with particular servers.
I've seen complaints that cross-server moderation doesn't work well in practice.
That is, a user deliberately joined a server with a moderation policy they liked. They see a lot of content that is supposed to be moderared out. They ask what's going on, and are told that content is being posted on other servers with admins with different policy views. This user is told they can either ban that specific user or ban the whole server. Neither is a good answer.
I don't know (and don't care enough to research) if this is a flaw in the protocol or a specific implementation.
Thanks for the explanation.
If a huge number of "laypeople" flocks to Mastodon, it's very likely that the users will be clustered around one or small number of "major" servers. So basically the admins of that server will become the de-facto moderators of discourse, it sounds like. Is that a reasonable course of events?
Or will the idea be more like "lefties should go to $SERVER1, righties to $SERVER2" and then the two servers don't federate with each other because they find each other's content distasteful?
According to this page, between September 1st and November 13th, the user count went from 6.12M to 7.56M, so an increase of 23%. Where does your estimation come from?
Even better would be a deployable binary service. Calling out all Go and Rust developers!