road to 1.0.0

[imrn99]

1.0.0

The crate is reaching a state (feature-wise) which I consider acceptable for a big release. This issue can be used to (a) argue about that claim, and (b) list in detail desirable stuff to do / process before such a release.

---

• Go over the public API / implementations / documentation

Because of (a) the size of the project, and (b) my code first rewrite later approach, the project probably has accumulated some unnecessary complexity. While I try to refactor my code once it works, it's not perfect each time, and it may have piled up. Going over everything looking for simplifications should be a good idea.

Done in #166 (doc) and ?

---

• Revise examples

There are too many examples, for not so much useful information. This is especially true for render examples.

Done in #156

---

• Revise public exports & multi-crate gestion

Seeing how bevy operates gave me idea for cleaner dependency specificication across our own crates. This, coupled with prelude modules, could give a simpler all-in-one crate to add as dep for using honeycomb.

This would also be an opportunity to formalize naming across the repo, since this is not a solved problem.

Done in #159

---

• Cleanup & test GRISUBAL

Tests have been successful using a total of ~30 2D geometries (see #151); including a trace of this in the main repository would be good, but I'm not sure how to do this without generating too much noise.

Done with #153, included in 0.5.0

---

• New render tool

Using wgpu was great to learn about render architecture / pipeline, but rewriting all graphical logic from scratch is not something that I wanted to do. In order to greatly enhance the visualizer (without greatly enhancing my workload), I rewrote it using bevy. See #144.

I'm thinking about making everything in its crate public to leave open the possibility of writing custom rendering code (e.g. curved edge making use of attributes).

---

• Check the dependency tree

With the visualizer's rewrite, the total number of deps seem to be approximately 400. I'm sure there's a lot that are not really necessary, most notably in optional bevy features.

Done in #165