A double coincidence of wants? I want to type. You want to read. I send to our relays. Relays' delete-timers start. Your client needs to pull from our relay(s) before deleted. The shorter the relay delete-timer, the lower the relay costs, and the more constant your client needs to be online and pulling. You wake up in the morning and read my notes, long after they've been deleted from our relays. So we'll either have to pay the relays, or have an always-on client? Something like that?

Reply to this note

Please Login to reply.

Discussion

The sender and/or the reader will need to pay for longevity, somehow, somewhere.

There is a heck of a lot of empty storage on remote servers, personal devices and USB sticks.

Idk, I only have about 1TB free to share ATM, which according to nostr:npub137c5pd8gmhhe0njtsgwjgunc5xjr2vmzvglkgqs5sjeh972gqqxqjak37w is practically useless

πŸ˜‚

1TB connected to a modem with v42bis 🐢🐾🀣🀣🀣

When desired, can a highly-followed-sender app encrypt their notes to each of their very many followers?

We already have private groups and we'll be getting private relays, and they have their own npub, so they have their own encryption key for all of the members, I guess.

It wouldn't need to be that short. Storage is cheap and json is tiny. I was thinking more like weeks on the very biggest, months on the mid-sized, years on the smaller, indefinitely on personal devices.

Datacenters currently store documents I wrote literally decades ago and I don't even know what all is in those folders and I don't care. πŸ€·β€β™€οΈ

Just deleting older versions of notes or notes that had been marked "delete" would help. Relays currently store EVERYTHING, even if the user specifically asked them to delete it.

Makes me think of Tahoe-LAFS.

Except that we have lots of HTTP servers with their own fileservers, and it's the client that usually handles encryption.