zlacker

I'm sure the loads are not comparable.

>>ameliu+(OP)
Equally, neither is the budget. Load isn't an excuse.

>>ameliu+(OP)
It withstands being one of the first link on HN, and it still loads comfortably.

For a personal website, that's a great performance.

>>ameliu+(OP)
Doesn't really matter if it works?

And if some massive org needs bandwidth for posts, can't they host their own public instance? I'm sure many organizations would prefer that over being at Twitter's mercy.

>>ameliu+(OP)
It's not crunching a bunch of analysis or ads in the background. I bet it scales wonderfully with maybe some kind of cache/CDN if necessary.

>>hinata+s1
HN traffic is not nearly the DDOS people make it out to be. I made it to spot #2 once and that was a few dozen hits per second at most. Maybe not something to serve with wordpress from a raspberry pi, but hardly apocalyptic levels of traffic.

>>WJW+S4
Wordpress on a raspberry pi will handle static, properly cached content just fine.

>>hinata+s1
previous HN articles linking to instances has caused them to become overloaded. this instance just seems to know how to handle it

>>ameliu+(OP)
Honestly.

what the fuck is this response?

I hear always that centralising everything is great because efficiencies of scale: but then we have something that works as good or better and the response is; “ah yeah, but the load is so high!”

Why do I care? I don't honestly give a shit about how much load you have, you could be factoring Pi on every page load; it means -nothing- to me. I kindly invite you to give more of a shit about user experience.

This also goes for when “complicated” systems fail, maybe making them so complicated and centralised is not the way.

>>dijit+ed
I've been arguing with Twitter emigres who seem to think that the UX on masto sucks.

I would say that Twitter is an automatic transmission, mastodon is a standard.

>>hinata+s1
HN itself is run on a single server. So long as you’re not running something bloated or trying to host on a DSL connection, you’ll be fine.

>>dijit+ed
Mastodon isn't truly decentralized. Every instance is its own feifdom and you have to communicate directly with them to exchange messages. This is unlike how Usenet, FidoNet, and SMTP work. They are truly decentralized as you can exchange data without directly accessing a peer's host server.

>>kevin_+We
I genuinely don't understand what you mean here. I don't directly access any peer's server when I want to talk to them.

>>kevin_+We
How do plan to send me an email without talking to my mail server?

>>dijit+ed
I couldn't see a few threads on Mastadon a few weeks ago, the post was up for 24 hours, and the perceived outage never dropped through that entire period, and probably longer after I stopped caring about the post. There were numerous people on HN experiencing the same.