I feel like part of the blame for the situation is that JavaScript has always lacked a standard library which contains the "atomic architecture" style packages. (A standard library wouldn't solve everything, of course.)
Edit: Removed a reference to node and bun.
https://developer.mozilla.org/en-US/docs/Web/API
The above seems fairly expansive, even if we remove all the experimental ones in the list.
Math.clamp is a big one (it’s a TC39 proposal). I’d also love to have the stats functions that Python has (geometric mean, median, etc.).
On the more ambitious end: CSV reading/writing and IPv4/IPv6 manipulation.
Deno's standard libary has nice CSV parsing/serializing, and you can use it in any environment.
Docs: https://docs.deno.com/examples/parsing_serializing_csv/
- range, clamp, inIvl, enumerate, topK
- groupBy (array to record), numeric / lexical array sorts
- seeded rng
- throttling
- attachDragListener (like d3's mousedown -> mousemove -> mouseup)
- Maps / Sets that accept non-primitive keys (ie custom hash)
- Functions that use map-like and set-like objects instead of maps and sets.
- Functions that should be async that aren't
- Async functions that should be cancellable, but aren't
- Functions that should use the disposable and async disposable stack (e.g. the `using` keyword)
- Functions that should return deeply immutable data structures but can't. (The R̶e̶c̶o̶r̶d̶s̶ ̶a̶n̶d̶ ̶T̶u̶p̶l̶e̶s̶ / Composites, and Structs proposals would help here)
- Concurrency coordination primitives (Imagine an array of 10 async functions that we wish to run. Going through them one at a time is too slow. `Promise.allSettled()` executes them all at the same time which might slow things down even more due to bottlenecking. We should have an easy way to say only execute a maximum of 3 at a time. As 1 function resolves, the next one is started, until all are resolved).