Proposal talk:Curve geometry

What is the point?

"Some map processing applications disregard nodes along a way and only include the first/last nodes. This tag allows these significant nodes to be identified and included. ​"

This is nonsensical. Even if they were renderers who "disregard" nodes they would also ignore this tag.--DaveF63 (talk) 00:10, 27 January 2020 (UTC)

Unclear title and purpose

Came in expecting a proposal on highway curve radius et al tagging. Was disappointed. Why don't you report such issues to the renderer? Kovposch (talk) 19:17, 26 January 2020 (UTC)

Because there's nothing wrong with the renderers. What is the purpose of this tag. The page doesn't state what it is.--DaveF63 (talk) 00:10, 27 January 2020 (UTC)

Based on the "curve geometry" nodes the proposer sprinkled around Cupertino, e.g. in changeset 77312037, I think the idea is to distinguish smooth nodes from corner nodes in a Bézier curve, as in the curve tool in Inkscape. At least that's how the tag often seems to be used. But tagging the smooth node by itself is meaningless unless the control points are also mapped. Control points would probably require an addition to the basic OSM data model, or at least some creative usage of relations, to say nothing of the challenge of getting data consumers to correctly represent Bézier curve geometries – not just renderers but also routers. This proposal needs a lot more thought before it could be a viable alternative to approximating a curve with a lot of corner nodes. – Minh Nguyễn ^💬 04:58, 9 March 2020 (UTC)

Just that based on my what little of my knowledge, it seems to me it would be more worthwhile and useful to tag physical curve properties on ways, a la the suggested railway:radius=* and abandoned turning_radius=*. Bézier curves don't directly relate to road and railway geometry. Splitting lines to make use of end-nodes would be less complicated than treating parametric curve. -- Kovposch (talk) 05:51, 9 March 2020 (UTC)