Hacker News Reader

Comment by whimsicalism | Hacker News Reader

They're wasting a massive opportunity. Mastodon should have been built to scale from the start because huge viral growth is exactly how social networks like this make it big.

As it stands right now, I am struggling to join because every mastodon instance I've checked is struggling with scaling.

View on HN

Mastodon isn't a commercial entity, so what's the opportunity that's being wasted? Its entire purpose is to provide an alternative model for social media on a non-profit basis. If that doesn't scale to Twitter or Meta levels so what? That's not the goal and it's certainly far beyond what those behind it could afford. Mastodon gGmbH has just two core developers.

Frankly, it's impressive that it's scaling as well as it has so far given the tiny size of the team and that all the servers are run on a volunteer basis.

Mastodon can scale, but it costs money. Most instances are hosted by people out of an interest to do good. Servers cost money... And for F/OSS project there aren't big pocket investors ... only funds, grants and hopefully donations.

Yes, when I say scalable, I mean ability to scale without having to spend bucketloads on new servers.

Many of the mastodon servers I am seeing pause new registration/break down are at a few thousand members, maybe 8k max. I dont know how Mastodon is implemented, but my intuition is that a well written C++, Rust, or even Java service should be able to handle that many users on a single instance.

Twitter makes -200m a year hosting their solution.

Mastodon by contrast spreads out this load to servers paid for by users, not by ads.

If you cannot join any servers, pay for a managed hosting provider to spin one up, or host yourself on a vps or a home server.

That may be well and true, but seems unrelated to the point i am making, which is that I would expect a single instance to be able to handle a few thousand users.

Sure one instance could support that if you pay amazon $2k/mo for a monster instance, but most do not have that kind of money to donate.

Most servers will be small, or you can host your own.

I think we are talking past each other.

I hear you.

It would be cool if someone rewrote it with a Crystal framework. Although there are limitations to this because it's faster than Ruby and similar to Ruby but not completely like Ruby.

https://news.ycombinator.com/item?id=33398253

It's Ruby on Rails, which is the second slowest web framework I would consider reasonable for production use (after Django).

Ah, it is all beginning to make sense now.

Or go with the best tool for the job - Erlang/Elixir.

Sure, 'ability to scale' (scalability) might be an issue, but if you are using it, you will see that the majority pf people are on other servers or communities, some not even using Mastodon. Mastodon itself is just one of the implementations of the federated protocol that uses ActivityPub. A lot of alternatives exists, like misskey, pleroma, etc.

The other opportunity "they" are missing is that normally when a new social network struggles under the weight of new users, that's good publicity. Instead of going that route, people get offered a huge list of niche or unavailable options, or told they should run their own server. And unsurprisingly a lot of people are saying, fuck that.

It's frustrating to watch. I guess a lot of fediverse people don't actually want twitter normies on board though.

The point of the fediverse was never to match the scale of Twitter on a single instance. In fact, that's something of an anti-pattern.

These are the inherent trade offs of a distributed social network.

> The point of the fediverse was never to match the scale of Twitter on a single instance.

Of course. I don't understand why you are telling me this.

Reply on news.ycombinator.com