[GeoJSON] Plug-Pull request to finalize a CRS-free I-D revision

[Howard Butler]

On Jun 13, 2013, at 6:48 PM, Allan Doyle <afdoyle at MIT.EDU> wrote:

> Final thoughts: I admit I was one of the people originally advocating dropping crs. I've changed my mind. We screwed it up, but it would be worse to compound the error.

Like Allan, I've changed my mind too. Changing srs in any way is likely to compound the problem. It is stinky for sure, and SRS interop beyond a simple EPSG code is likely to be more of a gentlemen's agreement rather than true interop, but its removal is not going to fix the problem. It will just fragment things and create more confusion.

I would like to propose that we add an Errata/Appendix C section to the bottom of the document that warns of the various pitfalls without making prescriptions about how people should act. Notes about srs would obviously go there, but also things like ring ordering in situations where spherical coordinates are being used, and axis ordering for the large swath of coordinate systems that GeoJSON insufficiently describes. This section can recognize the usage patterns that have emerged with GeoJSON in the 5 years since its release. Links to various mailing list discussions, github tickets, etc might also be useful for someone looking to do some deep research before declaring us and the specification to be idiotic.

Howard