Matthew Petroff

@Fil My projection has a closed-form inverse (that was part of the design requirements), but I never ported the Python implementation of the inverse to JavaScript, since I didn't need...

There are layout and scaling differences between the two implementations. The Python implementation directly returns coordinates in a quincuncial arrangement with $x, y \in [-1, 1]$. The JavaScript implementation returns...

I implemented the inverse in JavaScript (also CC0): https://bl.ocks.org/mpetroff/dcf7b090eabda85d081d47e8a0c71d4a Additionally, I updated the forward projection implementation (the `hexadecant()` function) to better match the paper. The copy in the supplemental information...

> how can I adjust the pop up? e.g.size etc? You can't using a `confirm()` (or `alert()`) modal. You'll have to write your own function to create and style an...

Very interesting. I think this is an edge case I missed when optimizing the tile display in 05dae19b2806fd01665351ab0f83a8bd203494ed. Before, a single tile resolution was used everywhere, which led to too...

I just fixed this in 950a2ce732cbe8d91f2b5483eaef175eee19675b. An edge case wasn't being handled properly.

> I didn't try hard to solve it, because no time. And you just assume I have time? I don't, since I didn't get a chance to look at this...

Someone emailed me about this a few days ago, but I haven't been able to reproduce it. That person fixed it by making this change:  It seems to just...

I'd definitely categorize it as a hack. I'm hoping the problem goes away with Chrome 59, since it seems to be due to a bug in Chrome.

I was finally able to find a machine on which I can reproduce this. Chrome seems to be throwing an erroneous `INVALID_ENUM` / code 1280 error on WebGL context creation,...