Corey Farwell

i'm down. should we open a separate issue for that?

I just went through the issues and labeled all the ones relevant to geo-types: https://github.com/georust/geo/labels/geo-types It would be good to go through all of these and decide whether they're worth...

I know this doesn't address all the points that @michaelkirk brought up, but one way we can reduce scope and potentially unblock some of this is by removing `M` altogether...

Regardless of _when_ we check validity of geometry, we'll need to implement the validity checks before that happens. Unless anyone else has a better suggestion, I think we can use...

For posterity, there was a discussion about validity of `Multi` geometry types, and we agreed that empty `Multi` types are considered valid https://github.com/georust/geo/issues/444

> > * Add a basic visual for each geometry type we offer ([Shapely does this](https://shapely.readthedocs.io/en/stable/manual.html#polygons)) > > > As for the how, here's the code for how Shapely generates...

This all sounds good to me. In regards to: > Concretely, we will have a FromWkt trait and a ToWkt trait for I/O on Geometry without knowing the exact type....

I created a couple extremely basic dependency diagrams to illustrate what I'm talking about. --- What @sunng87 proposed  --- What I counterproposed 

@georust/admins This is a pretty significant architecture/design decision. Would be great to get everyone's feedback

Right now, I don't feel strongly one way or the other, but the biggest reason I don't like `ToWkt`/`FromWkt` is it implies that `rust-geo` will have to implement every single...