> Self-describing is point-less for serializations
you couldn't be more wrong. what happens when you lose the schema, or never had access to it in the first place? think from the point of view of reverse engineering
I imagine when someone is choosing a serialization solution they probably don't really care about people trying to reverse engineer it... And if they did, they would just make schemas available instead.
And if you lose your own schema, then you probably have more serious underlying problems.
All the binary JSON schemes, like CBOR for example, are all about making encoding and decoding faster. So the world kinda wants this even if you don't.
Do I want protobufs replacing JSON? No. I want flatbufs augmenting JSON (and XML, and ...).
I maintain an ASN.1 compiler and run-time that supports transliteration of DER to JSON. This is possible because the syntax/schema language/IDL and the encoding rules are separable. This is how you get the best of both worlds. You can use optimized binary encoding rules for interchange but convert to/from JSON/XML/whatever as needed for inspection.
you're not getting it. imagine if every single public website today using JSON switches to protobuf, you instantly lost most if not all ability to reverse those sites.
All these sites go through an evolution where they start out having open APIs, attracting third party developers, then eating their lunch and closing the APIs. No one in this sub-thread is saying that's desirable. But also nothing you can do can stop them from using something other than JSON.
And if you lose your own schema, then you probably have more serious underlying problems.