I searched around for observablehq. I hadn't see it before. Checkout https://observablehq.com/@observablehq/five-minute-introduct... and https://news.ycombinator.com/item?id=20184181.

I wonder how many of those would be feasible in real life situations

For example, a celestial body coming in proximity to another will impart tidal forces, might involve some kind of mass exchange or simply find form of atmospheric drag.

The three-body problem has very few stable solutions, which include an equilateral triangle and a figure eight:

https://sites.math.washington.edu/~morrow/336_12/papers/adri...

"The actual probability of finding such a [figure eight] star system is somewhere between one per galaxy and one per universe."

This is an extremely misleading statement. The Earth-Sun-Moon is a 3 body system and quite stable. Obviously moons and multi-planet systems are quite common. As are 3-star systems. Our nearest neighbor is one: Proxima/Alpha Centauri. And Proxima Centauri has a planet too.

What is difficult to find stable solutions for is 3 body movements where the three objects are of similar mass. But that is an improbable circumstance to start with.

Also that citation seems to be totally disregarding the fact that multi-body systems never start in stable systems but instead naturally align into them.

If I work on an observable notebook, is it (or can it be) backed by a git repo? Is it convenient to edit the code in a traditional text editor?

Not too easily, and no.

With Observable notebooks we're trying to expand the boundaries of how live, networked and interactive a notebook can be. So, to that end, the notebook is the editor, visualizations update reactively with your changes to the code just as they would update to changes in the data, you can inspect and autocomplete actual values flowing through your program, and you can fork someone else's notebook or merge their changes with a single click.

There's built-in history too (live, every version reproducible), as seen in 15 seconds with an animated gif here: https://observablehq.com/@observablehq/history

In the future, we're thinking about working on APIs for better git and generic text editor integration; but for the reasons listed above, the primary focus for now is on making working directly in a notebook on the web the best experience possible.

Thanks! I have another question about observable: are you aware of anyone using mathbox in conjunction with observable? Do you think that could be a fruitful direction?

https://github.com/unconed/mathbox

I'm not aware of anyone using Mathbox in Observable, although it certainly seems promising!

On the other hand, I have seen quite a lot of neat Ganja.js notebooks float by, for plotting explorations of algebraic spaces in notebooks: https://observablehq.com/search?query=ganja

But do these 'analytically' orbit? I reckon such a proof must be quite intricate? Are there proofs for each of those families?

I googled the name of the first set of orbits and found this https://arxiv.org/pdf/1303.0181.pdf

Routine plea: please link to abstracts (Šuvakov and Dmitrašinović - Three classes of Newtonian three-body planar periodic orbits - https://arxiv.org/abs/1303.0181) rather than directly to PDFs. (It is an easy click to get from abstract to PDF, but not vice versa.)

That's good advice. Thanks! I've switched a couple of the links to abstracts instead of PDFs, though what I really need is a proper references section since direct links are also convenient—if they work.

They're interesting, hypnotic, and beautiful!

This notebook sharing website look nice. Why we haven't seen this before?

Previous submissions may be of interest: https://news.ycombinator.com/from?site=observablehq.com