# Re-imagining nostr.directory
https://nostr.directory was a good and much needed idea for bootstrapping nostr and contact discovery but had a few things wrong with it:
- centralized - required maintenance, data, backups, processing
- spof - has built-in single point of failure, me as the bottleneck when I get distracted by life it stopped being useful
- raw data ≠ useful data - even though it was a good attempt to help people connect their online identity together, that data is useless unless it can be utilized by apps and users in a helpful way.
**Next Steps**
As pondered and discussed with nostr devs and users many times, nostr.directory can have and serve connection and identity data of each user, next much needed step is an actually usable, easy to understand “reputation score” to help nostr clients and microapps surface relevant content and weed out spam.
Important tenants of such a system will be:
- Opt-in: If a users does NOT want that “for you” page they don’t have to see it and use it.
- High Availability: once a user wants to curate their microapp using WoT data, it should be trivial for apps to implement it.
- Nostr Native: Completely open-source and self-contained repo including all processing, backups + nostr relay serving available data.
- Context Based Scores: Having a 0-100 reputation score is nice and clean but it’s also important to know each separate field of expertise can come with a “field specific score” which is to say, I may value npubx’s views on bitcoin related things but not on latest diet trends.
For this I’ll do some research to existing methods and put an early version out there and try to integrate it into existing apps maybe even write a few demo apps that utilize it. Will go from there.
All inputs are welcome and appreciated:
pseudozach (npub1xhe…56yf)
hello@nostr.directory
https://twitter.com/citlayik
relevant links & posts & prior work & references
WoT post on SN: https://stacker.news/items/477386
how stacker news does it: https://stacker.news/items/477026?commentId=477086
hivemind: https://hivemind.vc/identity/
noswot: https://codeberg.org/weex/noswot
tools:
https://github.com/nostr-dev-kit/ndk
npub1xh…m56yf
2024-05-06 21:47:11
Author Public Key
npub1xhe9408d5hm3dpwax78sy9nuc5warycnvcy4r3qzv6jacwu26r6s5m56yfPublished at
2024-05-06 21:47:11Event JSON
{
"id": "8ae88aa6f130935c4bb782a715dec5440855d991eb4b2ddbeba1f1978067d9a4",
"pubkey": "35f25abceda5f71685dd378f02167cc51dd19313660951c40266a5dc3b8ad0f5",
"created_at": 1715032031,
"kind": 1,
"tags": [
[
"p",
"35f25abceda5f71685dd378f02167cc51dd19313660951c40266a5dc3b8ad0f5"
]
],
"content": "# Re-imagining nostr.directory\n\nhttps://nostr.directory was a good and much needed idea for bootstrapping nostr and contact discovery but had a few things wrong with it:\n\n- centralized - required maintenance, data, backups, processing\n- spof - has built-in single point of failure, me as the bottleneck when I get distracted by life it stopped being useful\n- raw data ≠ useful data - even though it was a good attempt to help people connect their online identity together, that data is useless unless it can be utilized by apps and users in a helpful way.\n\n**Next Steps**\n\nAs pondered and discussed with nostr devs and users many times, nostr.directory can have and serve connection and identity data of each user, next much needed step is an actually usable, easy to understand “reputation score” to help nostr clients and microapps surface relevant content and weed out spam.\n\nImportant tenants of such a system will be:\n\n- Opt-in: If a users does NOT want that “for you” page they don’t have to see it and use it.\n- High Availability: once a user wants to curate their microapp using WoT data, it should be trivial for apps to implement it.\n- Nostr Native: Completely open-source and self-contained repo including all processing, backups + nostr relay serving available data.\n- Context Based Scores: Having a 0-100 reputation score is nice and clean but it’s also important to know each separate field of expertise can come with a “field specific score” which is to say, I may value npubx’s views on bitcoin related things but not on latest diet trends.\n\nFor this I’ll do some research to existing methods and put an early version out there and try to integrate it into existing apps maybe even write a few demo apps that utilize it. Will go from there.\n\nAll inputs are welcome and appreciated:\nnostr:npub1xhe9408d5hm3dpwax78sy9nuc5warycnvcy4r3qzv6jacwu26r6s5m56yf\nhello@nostr.directory\nhttps://twitter.com/citlayik\n\nrelevant links \u0026 posts \u0026 prior work \u0026 references\nWoT post on SN: https://stacker.news/items/477386\nhow stacker news does it: https://stacker.news/items/477026?commentId=477086\nhivemind: https://hivemind.vc/identity/\nnoswot: https://codeberg.org/weex/noswot\n\ntools:\nhttps://github.com/nostr-dev-kit/ndk",
"sig": "35d4ef006076d9085bc74196b84de512494b9ccc799b0c5984b1f4fc4eb43965cbc4445cd51fc520cded5944478bc170e7e84ca4bbff28e75606bb0265ceb425"
}