Michael Kirk

> How' this revised initial plan sound: start measuring individual events (wide road crossing for pedestrians, many-legged intersection crossing for bikes+peds, car-wants-to-overtake-bike for bikes), display them for individual trips, and...

As @rmanoka alluded, I think that the new test is invalid because polygons don't `contain` points on their boundary. If it did, there'd be no difference between `contains` and `intersects`....

> would it be possible to add a robust convex hull algorithm? Good question. I'm not sure if it's currently robust or not. I know that we're using Kernel::orient2d, which...

To consolidate some conversation that is happening both here and in discord: > how does that relate to your original issue @frewsxcv noticed in rgis that when using our concave...

> What if the trait implementer wants to control the scalar values in the point more tightly? Specifically, something like rounding an f64 to a certain number of decimal places....

Just to make sure I understand your situation, in the example you're showing, you ultimately want to create a polygon whose exterior will have overlapping segments, and is thus "invalid"...

I think this is clever and is in pursuit of solving a real problem that I have. I can see how the approach you laid out could be useful in...

I saw some modest improvements (~5-10%) to serialization/deserialization with [my tinyvec branch](https://github.com/michaelkirk/rust-geojson/tree/mkirk/tiny-vec). And much bigger improvements (78%) with strictly geometry parsing, which I suppose makes sense, since this only affects...

Some related issues from pelias: **include geometry in `/places` response** https://github.com/pelias/whosonfirst/issues/19#issuecomment-370545690 **create new endpoint for geometries of cities** https://github.com/pelias/api/issues/1121

Maybe this is a dupe of #69?