Default avatar
npub1dww6...pdnf
npub1dww6...pdnf
Wether Nostr is struggling or people who develop on Nostr are struggling, it does not matter. Nostr solves the problem of distributing immutable unencrypted textual data in a decentralized, censorship-resistant way, aka kind1. I believe it’s the best design for this kind of problems. However, people try to solve other problems that are either or all of 1. Encrypted data (DMs) 2. Non-textual / multi-media data (images / videos / streaming) 3. Mutable data (configurations, lists, collections) Then we see an ocean of ineffective hacks. A better approach for an application vendor, indie or group, open or close sourced, solo or funded, is to just use Nostr for its strength and implement other parts in more suitable ways. Nostr does not need to become the carrier of everything. If it does one thing extremely well, it will flourish. View quoted note →
As much as I admire Evan Vue’s effort and creativity of inventing Vite. I consider is not an essential software for the web development and in fact it introduced unnecessary complexity. If build/bundling is essential, then esbuild is more essential than vite. Of course, bundling should never be essential. It is artificial.
Is NIP-46 always connected via a Relay? Can the client and the remote signer connect directly via a http request?
2 months ago @zoé recommended me this book. On my trip to Thailand, I accidentally walked into a bookstore with English books, found and bought it. I never opened it, coming back from my trips. Yesterday night I had a bad sleep and felt miserable today so that I finally started to read it. I love it and will finish it in the coming weeks. image
@reya One problem Blowater has is that it heavily uses IndexedDB which is slow and has horrible APIs. As much as I can optimize it, it’s a bad foundation to be built on. I believe your investment in native desktop technology is very worth it in the long run. Web is very good at presenting / displaying things but very bad at everything else. Tauri is a good approach. We can use Rust to implement a performant storage layer. I look forward to see what you make with Coop
20K lines of code is the sweet spot that a good developer can realistically hold in his brain at once per year. Which means, the code size should be Total Lines of Code / (People - 1) <= 20K You always want to -1 because you want shared knowledge about the codebase. If your team is suffering from complexity but the number is <= 20K, you designed the wrong architecture.
Plain text is the best UI human has invented so far. It’s insanely effective. That’s why chat apps are so powerful. You can arrange meetings, events and complex tasks in the chat, without using a specialized app. Chat is the app of everything.
I really enjoy the 1 hour before sleep. I feel the complete ownership of this hour.