JuliaManifolds / ManifoldDiff.jl

Differentiation on manifolds
https://juliamanifolds.github.io/ManifoldDiff.jl/
MIT License
10 stars 2 forks source link

Proposal / Discussion Logo #44

Closed kellertuer closed 1 week ago

kellertuer commented 4 weeks ago

I had an idea a while back to try to get our packages a bit better distinguished by creating a few more logos; maybe we could also ask curmullion for help to improve the logos some time.

Here is a first sketch of the generic idea: Take the Manifolds.jl logo and replace one (or two) patches by interface patches (to make them calmer) and place something on top. Here I was not super creative.

But for DiffEq one could replace the top and instead of the patch draw a solution of an ODE there as a green curve. here since it is text and the $\nabla$ usually on the left, I replaced the red one. In this experiment I also tried blue geodesics, but we can check what best to do there otherwise.

logo-diff-dark logo-diff

mateuszbaran commented 4 weeks ago

Sure, we could add logos to more packages. In this case I don't think the $\nabla$ sign fits there but I don't really have a better idea.

kellertuer commented 4 weeks ago

we can also take a $\partial$ ? Maybe that fits better.

I think the blue geodesics are nice, but I would be open to also further use white/black there.

kellertuer commented 4 weeks ago

I indeed like the partial version better:

logo-diff2

one could of course also do the partial even larger so that it reaches further atop and to the right (maybe lets not increase it to touch green or violet, but a bit):

logo-diff3

Then I tried to maybe leave out the “interface patch”, but that feels to me a bit unbalanced (and I miss the red color):

logo-diff4

So I think I like the second (with interface, large letter) best.

mateuszbaran commented 4 weeks ago

Yes, it looks better with the partial derivative, good idea. I like both the first and second version with $\partial$, so if you prefer the second one, it's fine for me :slightly_smiling_face:

kellertuer commented 4 weeks ago

Nice. Will do that the next days when the main organisations for the semester start (it starts tomorrow) are done.

I could do the similar approach for DiffEq, but would need a curve “whirling” around the North Pole for that from some ODE.

mateuszbaran commented 4 weeks ago

Maybe the example from https://juliamanifolds.github.io/ManifoldDiffEq.jl/stable/examples.html would be fine? IIRC it whirls somewhat if you set a large time step with a low order method.

kellertuer commented 4 weeks ago

I can check, a problem might be, that it either has a thin line with and is not easy do see or looks a bit too clunky. But maybe a good start to implement loading data and pdf-plotting it.

kellertuer commented 3 weeks ago

In principle I set up the documentation to use the new logo. In practice, I forgot how to generate the png with half-transparency for the geodesics, have to check how I tricked that back when we did the first logo.

kellertuer commented 1 week ago

I finally found the time to tweak the logo in GIMP to have a transparent background both in light and dark mode, now the docs and the favicon look nice. so this is finished to go. Should we release that as a new version or just put on master and release it with the next changes?

mateuszbaran commented 1 week ago

Cool! There is no other change being worked on here right now so we can make a new version just with the logo.

kellertuer commented 1 week ago

Ok, then we do that.