mleku on Nostr: DVMs with specialised graph databases feeding purely from follow lists would be the ...
DVMs with specialised graph databases feeding purely from follow lists would be the way to make it faster
i know with the DB i used, derived from fiatjaf's badger implementation in eventstore, each npub can have a simple 64 bit unique conflict free identifier, and building graphs out of that is pretty cheap actually
Published at
2024-06-25 18:04:23Event JSON
{
"id": "84acfd2c87e9c8a33a2817446d2f3b973b51a00e87cac8f2378ff7d0dc0315fa",
"pubkey": "4c800257a588a82849d049817c2bdaad984b25a45ad9f6dad66e47d3b47e3b2f",
"created_at": 1719331463,
"kind": 1,
"tags": [
[
"e",
"9db0ce842237302996d8cf68ee2413cbcf4b2ca72b57763e397ba9fd4b15c3db",
"wss://nos.lol/",
"root"
],
[
"e",
"300dde277918dddf737c881f89efcd5ca1f64fe66e8e911dcc15a00c61366316",
"",
"reply"
],
[
"p",
"e5272de914bd301755c439b88e6959a43c9d2664831f093c51e9c799a16a102f",
"",
"mention"
]
],
"content": "DVMs with specialised graph databases feeding purely from follow lists would be the way to make it faster\n\ni know with the DB i used, derived from fiatjaf's badger implementation in eventstore, each npub can have a simple 64 bit unique conflict free identifier, and building graphs out of that is pretty cheap actually",
"sig": "47ddb13d6e686253ec8ea2593e1c20357afdb8cdca3668bd89a7105e61c7ee7102029be5a5d5005fcdd8fbcb1da55630e66b28b36a9a8dd90485ffe9ce2bd5aa"
}