@rauschma there are two minor changes to the spec which can make the feed content-addressable and reduce duplications (when we have the same article from different feeds): 1. author public key and digital signature, 2. trusted timestamps https://en.wikipedia.org/wiki/Trusted_timestamping
@nzakas thank you! Good to know. I always have a problem with the NaN name, I would better call it undefinedNumber, like in math. But, we can't change IEEE 754. I guess.
@rauschma I love TypeScript but I don't like additional build steps. IMHO, JavaScript with JSDoc comments and TypeScript as a linter is a much simpler approach. Hopefully, we will see JavaScript type annotations one day.
@rauschma there is another reason you may prefer lazy concatenation (version 2): optimization. For example, in theory, `console.log(largeText, anotherlargeText) should work faster and consume less memory than `console.log(`${bigText}${anotherBigText}`) since we don't need to create a huge temporary string. In FunctionalScript, I use lazy lists to describe HTML, and if I really need a string, I can call a `htmlString` function: https://github.com/functionalscript/functionalscript/blob/main/html/module.f.cjs#L116
@rauschma IMHO, TypeScript provides good support for algebraic types: ``` type Product<A, B> = readonly[A, B]; type Sum<A, B> = readonly['a', A]|readonly['b', B]; ```
@rauschma I made a rule for myself: avoid user defined nominal types, such as JS classes and symbols. Nominal types identified by location (where they are defined) so it makes serialization/cloning almost impossible or very hard. So, always structural types and option # 2.
@Migueldeicaza, I can imagine how discouraging such comments "just google... expect to fly..." can be for people without huge experience in software development like you have.
@laund I think it's a feature, not a bug. If Rust can find an object with the name then it will use it instead of declaring a new one. I wish Rust has something like `let ANSWER_TO_QUESTION => ...` to make sure that we would like to create a new object.
@cwebber we should add decentralized identity to our blog site and digital signatures with trusted time stamps to our posts. Then we can build a lot of cool stuff (including social networks) and own our data and identity.
@KirillOsenkov I wish we didn't depend on centralized identifiers so we can always own our content, contacts, and followers. Sadly, Mastodon has the same problem.