And like yeah, both the wonderful (and foss!) .json5 and Microsoft’s semi-proprietary(?) .jsonc exist, but most projects just use their language’s default JSON parser that doesn’t recognize them. What I would personally love to see is .json5 support baked into the default JSON parsing libraries of Python, Go, etc. (Enabled by a flag, likely.) It’s a superset of regular JSON and fully ES2019 compatible, so there shouldn’t be any issues.
I would add PairDrop to your list to have bookmarked. It’s completely web-based so no download required and thus fully cross-platform. It also works across different networks (i.e. over the internet) by pairing devices or creating a room. Basically Apple AirDrop, but universal and on steroids.
Eh, I’m not sure if the world can be picked backup, it’s been knocked down pretty hard. I’m always down for trying though.
/apr1
Except in very rare configurations (i.e. not 99.9% of residential), you do not want to have multiple paths to ground within a system. All grounds should go to the tied ground/neutral bus in the main breaker panel, which then goes to earth via a ground rod or a clamp to a copper gas/water line, etc. Otherwise you can have current flowing in ways that the system isn’t designed for, which at the least can trip breakers and GFCIs, and at worst exceed the rating of the wires in a short condition and cause a fire.
Except in very rare configurations (i.e. not 99.9% of residential), you do not want to have multiple paths to ground within a system. All grounds should go to the tied ground/neutral bus in the main breaker panel, which then goes to earth via a ground rod or a clamp to a copper gas/water line, etc. Otherwise you can have current flowing in ways that the system isn’t designed for, which at the least can trip breakers and GFCIs, and at worst exceed the rating of the wires in a short condition and cause a fire.
For small personal deployments, is SQLite support planned? It’s crazy performant and I have to imagine it would work for up to 500 contacts at the very least, which should cover the majority of deployments. Making Redis optional (otherwise using a basic in-memory KV store of some kind) would also be cool.