Just doing my bit to ensure that some of that conversation gets archived. Nothing to see here - move along.
#JSON came on the scene and quickly adopted an attitude of "I alone can fix it". Then propagated a bunch of populist myths - #JSON is smaller than #XML (debunked on their own wikipedia page) XML is complex (as if you are compelled to use all of it for everything) and ignoring the fact that much of that complexity derived from the domains it was deployed in.
Everything is justified by "It's simple and lightweight" which is great except they create this expectation that it be deployed for things that are not simple or lightweight.
This leads to the "Nobody could have thought that ...... could be so complicated" and the realisation that you do need schemas, namespaces (i.e all the things XML was condemed for).
However a spirit of "Repeal and Replace" pervades there are no "Good Parts" in anything the "other side" did you end up coming up with your own version , which when scrutinised is
a) found to cater for the interests of one community (Developers) at the expense of every one else. and b) founded on false equivalences - (JSON Pointer = #XPath) and not to sing the praises of #XMLSchema #XSD but try ripping it out and replace it with JSON Schema and see what happens.