2025:
Besides working on Logical Map and refining Floppy PNG, I learned about many forms of messaging. I sent map updates over #ATProto, #Nostr, and ended up with my own modified pure WebSocket version using the Nostr (NIP-01) event object for extensibility. The conceived application was for collaborative knowledge graphs. I verified each update with a signature, running a Deno server to process. In the end, I decided that it was too much to add to my main paper, which I'm still writing, and cluttered up the application code. Instead, I focused on merge features and import/export in the app. I'm happy with this approach.
2026:
I need to finish the Logical Map paper. I'm targeting the complexity and format of the Bitcoin whitepaper. (I'm really nerding out on the format consistency for both PDF and HTML versions). I also want to build my own messaging infrastructure based on what I've learned in 2025. My current thought is it should be a pure Mosquitto MQTT broker with Client TLS self-signed certs. I'd like to integrate with other feeds like Telegraf. There is a sweet spot somewhere of metrics/monitors and collaboration on maps as systems fail, and I think I can add some perspective there with my experience running large IT networks. I think it will be useful to use the Nostr event object @kind 1. I appreciate the beauty of that. But, I also think that securing the identity of the client and server is important for spam/security/cost perspectives. I'm fairly sure I won't double up. That is, I won't use secp256k1 sigs if I'm already using x.509. But, I'll do a decent analysis of that and wait before I pin a solution. This *might* also serve as simply a secure tunnel/messaging. That is, as long as I have the Nostr event object, I could treat the MQTT/Client TLS as a tunnel that passes messages with proper secp256k1 sigs. This *might* also serve as the desired key delegation mechanism. The server could filter npubs and route to MQTT topics. Well... that is enough for 2026. The big priority is Logical Map, of course.
