Analogue Dog's avatar
Analogue Dog 2 weeks ago
Interesting. Thanks for explaining. I'm not sure I understand what you mean by web [for lightning] though. I get your point about synchronisity.I haven't tested mobile lightning wallets in anger for a couple of years now, but assuming that receiving payments while the wallet app is backgrounded is still somewhat hit-or-miss across mobile OS's / versions / custom roms etc. My favoured model at the moment is use of Ed25519 as a user profile layer, and using PKARR for servers, payment details etc. This allows users to designate a number of Bolt12 (or whatever) addresses in order of preference... so perhaps synchronous as 1st, CoinOS as 2nd. Or even mw as 1st, monero as 2nd etc.... early beta just dropped today.

Replies (4)

People interact via the web, http/https, so for lightning to be the money of the web it needs to work on the web. Fortunately Nostr picks up where lightning failed since it uses http and can get invoices there. CLINK exists for this reason. Lightning having its own p2p layer necessitates native apps on mobile, stands out on militarized networks, and couples communication and monetary identity. Pubky has a few high level concepts right directionally but relies on DHT which doesn't work, after 20 years the torrent world still uses trackers because DHT is so slow and unreliable. demonstrates static pages just using nostr to fetch invoices, no galaxy brain p2p required.
Analogue Dog's avatar
Analogue Dog 2 weeks ago
I suppose the gatekeers of whether or not lightning works on the web, are browser dev teams. Easily fixed but if I'm not mistaken some of the biggest antitrust cases have been about browser distribution. Mainline DHT may be slow, but PKARR's can be cached and propogaged P2P between pubky home servers (with TTL of course). Private DHT's could also be used. I couldn't get the pay invoice to work. But given that most things on Nostr don't work, no big surprise I guess. image
John Carvalho's avatar
John Carvalho 2 weeks ago
PKDNS responses from Mainline DHT are actually extremely fast, especially if you don't need to wait for the net-wide response.
John Carvalho's avatar
John Carvalho 2 weeks ago
Your claim of DHT being unworkable and slow is incorrect. Go ahead and set up PKDNS / pkarr relay and try it. DHT is not used for general searches, only for discovery of endpoints for specific keys. “Torrent trackers still exist” is not an argument against DHTs. It is an argument for layered systems. Trackers are convenience infrastructure. DHT is exit infrastructure. Pubky uses the DHT for the root of discovery, not for the hot path. That is why it is fast. Resolve the signed record, then use the web (or any network, really). Nostr relays are the tracker analogy here, not sovereign. Pubky’s point is that the key remains the authority. Also, your “Nostr uses HTTP” framing is sloppy. Nostr’s base protocol is client-to-relay over WebSocket, not plain static HTTP. That does not make it bad, but it matters if you are criticizing one network dependency while actually depending on another anyway...