MikeBeas · HackerTrans

7 時間前議論

Bluesky is not all of atproto, first of all. There are a bunch of apps thats don’t use any relay at all. Second, even if you aren’t on bsky’s relay, there are others, like the Blacksky relay.

I’m guessing from your constantly-incorrect capitalization of “BlueSky” that you’re not familiar with any of this even a little bit and you’re just repeating what you’ve heard from other, equally unfamiliar people.

MikeBeas

8 時間前議論

Relays are trivial to run now with changes to how syncing works a while back. AppViews are entirely up to the designer of the view. You can make it as complex and expensive or cheap and fast as you want, it’s entirely up to you. You do not have to provide the exact same functionality as Bluesky’s AppView and can basically do whatever you want.

MikeBeas

8 時間前議論

Tons of people have started products without issue. But if you flood a popular relay with data that exceeds its rate limit then getting throttled is to be expected. This doesn’t prevent your participation in the protocol or network.

There are other relays and indexes out there that may have different rate limits. That’s just how these things work. The data from W is still there on their PDS, still public, still crawlable, and able to be backfilled within rate limits.

This is a non-issue.

MikeBeas

8 時間前議論

My Mastodon instance blocked Threads preemptively because they were afraid of the traffic driving up their hosting costs so you’re basically just very wrong or lying about this. A huge flood of incoming traffic can be hugely expensive for ActivityPub node operators. The problem of giant onboarding events is real, there’s just no default protection against them in Mastodon like there is in Bluesky, so operators either have to take matters into their own hands with defederation or risk driving their bill through the roof.

コメント