Tim Schaub

GeoJSON does not require that geometries have only 2 or 3 dimensions. It recommends avoiding more than 3, but doesn't forbid it (see the "SHOULD NOT" in [Section 3.1.1.](https://www.rfc-editor.org/rfc/rfc7946#section-3.1.1)). In...

> We only allow 2D or 3D geometries, at the moment (although that is a bit hidden in the description of the "encoding" keyword) I completely missed that.

I agree it makes sense to support deeply nested feature property values. I'm not sure that the `nested` expression is really the best way to do this. I think the...

The thing I'm not clear about with the type hint is where it is actually required. I understand how it works currently, but am uncertain where it is required. When...

Thanks for the proposed changes, @Dolpic. I don't have time right now to invest in this, but I think it would be ideal if we could come up with a...

In the example above, the performance issue that I see has to do with the number of `Image` elements being created (one per point feature). You can see similarly bad...

If you can include a link to the GeoTIFF you are using, that would make it possible to debug.

@jahow - Thanks for taking this on. Having more complex / realistic style examples will really help identify what else needs work. I'm curious about the dynamic variable name part....

I see now that the dynamic `var` is about the icon offset variables. As an alternative to this ```js { 'icon-offset': ['var', ['concat', 'offset-', ['get', 'maki']]], } ``` we could...

It may be that `between` is not really the most useful operator here. It sounds like you want a `case` operator to evaluate a number of boolean conditions, applying a...