alexburner / dimension-trails

Orbital motion through 01234D & time

dimension trails

Orbital motion through 01234D & time

Static site — https://alexburner.github.io/dimension-trails

Playground — https://alexburner.github.io/dimensions

Development

install

Install npm dependencies

npm install

dev

Run dev server + watch (outputs to dev)

npm run dev
npm run clean # remove .cache/ dev/

tsc

Run TypeScript compiler + watch (no emit, types only)

npm run tsc

format

Run prettier and tslint --fix

npm run format # prettier + tslint
npm run prettier
npm run tslint

build

Run production build (outputs to docs)

npm run build

dimension-trails's Issues

Use JSON.stringify/parse for worker messages

This is a faster method than the StructuredClone algorithm [ insert evidence here ]

However, TypedArrays don't play nicely with stringification [ insert evidence here ]

So, some sort of conversion would need to occur with the Vectors being exchanged between contexts. Or at least, appropriate types for this intermediary state, and proper handling on both ends.

• the main thread converts Particle objects to Particle3 anyways, and indexes into the Vectors directly, so it is already (accidentally) compatible

• the worker thread will need to convert incoming Particle objects in the 'init' message handler (luckily, this is not a hot path). Possibly a while loop, and break when an index is undefined?

const v = new Float32Array()
let i = 0
while(obj[i] !== undefined) v[i] = obj[i]
return v

Add new radialRandomVectorFromOther(length, otherVector)

New radialRandomVector() function that:

• preserves negatives from other vector (since they get lost in the squares)
• preserves magnitude (radius) of other vector
• preserves existing values of other vector (so, kind of a radial random fill?)
• handles mismatched dimension counts (main use case: lower → higher)

Also for the love of all that is good, is there a better name than radialRandomVectorFromOther()

This should improve consistency when generating initial particle sets from low to high dimensions