Yes, though I'd say a whole protocol to transport nostr events is probably better defined seperately as it touches a different layer than nostr itself is on. Such a protocol = layer 4 (of the OSI model) Nostr = layer 7

A protocol that allows everyone to become an ISP and earn money reselling their internet connection. Tollgate.me

If by external you mean public. Then no. Tollgate networks are fully natter, meaning no public ip's

ah interesting! I think btw if FIPS works out (replacing IP with pubkey routing) then that would create a huge opportunity to clean up higher layers that are now bothered with encryption. Encryption negotiation/handshakes can then be done on the same layer as the IP layer, removing the need for things like SSL/TLS altogether.

TollGate, NoDNS, FIPS, #soveng, SugarDaddy.Cash

Correct. In the current state of the code automatically inserting the certificate is still VERY risky because I haven't implemented certificate security checks yet. If the checks are not in place. any [npub].nostr could publish a self-signed certificate with *.google.com and your system would trust it. Allowing a MITM attack. Just be aware of this when testing. It's very experimental.

Thinking about it. Transmitting Nostr events over WebSockets doesn't make much sense from a networking perspective. It's quite inefficient... TCP has checksums to guarantee perfect transmission and always acks every transmission. And websockets live on top of that, meaning more complexity/delays. 1) In Nostr events, the checksum is already in the event itself (id/signature) 2) Nostr is stream based (more like udp), we don't have to ack everything. 3) Nostr events have a standard format, meaning you can strip off the json during transport. So in layer terms: WebSocket -> TCP -> IP -> PHYSICAL becomes 'NoStride' -> IP -> PHYSICAL In a fully nostrified world, i can see a protocol for JUST nostr subscriptions be justifiable. K, GN

Flash your Tollgate Release Explorer: https://releases.tollgate.me/ Signal Deployment Group: https://signal.group/#CjQKIFUj7wFVxIbHjujYAWGTKeMIvi1DdVYk2Zem3uNlmIwWEhAiFlw7arUvnitB1V0_cTaA

Interesting deep dive into TCP https://www.youtube.com/watch?v=rmFX1V49K8U

You can find the repo here: https://gitworkshop.dev//relay.ngit.dev/no-dns It contains the spec, a cli to publish the records and the DNS server

'Boox' seems to be quite good. As long as you go with an Android tablet ... Of course the Daylight, I love that it can do pretty much anything but does come with a hefty price tag. Writing on it is very comfortable

NoDNS in action 🔥

NIP-01 chose *A* transport method that works very well and will work well for a long time. It didn't have to define a transport, because really any transport works. Which is nice because it allows us to play with the layers below. Websockets in high throughput it will suffer the inefficiencies. Prob the current-day hangup is still checking signatures. So i'm not arguing we need this today, it's just a cool possible optimisation later on.

Airlines in warm destinations: "Now let us spray this totally harmless substance in this closed off tube. Just cover your eyes, mouth, food and drinks when we spray it, as it might burn. It's for your health...."

Dehumidify your tech stack