> Standards are good for everyone, even when very old and seemingly outdated.
This is definitely not the case. As an example, the Official Standard way to encode data structures in the 90s was ASN.1 from the ITU. Problem is, ASN.1 kinda sucked and so nobody uses it anymore outside of cryptography. Then the Official Standard became XML. Eventually people decided that also wasn't that great and switched to JSON, which is a totally unofficial standard defined by one guy.
Your approach / the EU approach would mean REST APIs would all need to be encoded with ASN.1 or XML.
This is more evidence that engineering for every complicated case will give you a complicated standard. And note that most of the complaints are often retread with arguable advances later.
JSON is still an odd one. 90% of that standard is "javascript, but only numbers and strings are allowed." With some extra shitty rules around comments and no real specification for what numbers do in the host language.
ASN.1 is good though. What sucked was everyone implementing their own serde tooling. In another, better world we'd have consolidated around ASN.1 DER and not wasted all that time and energy with XML, JSON and so on.
If it was really that good there'd be plenty of great open source implementations and tools by now, but in reality nobody wants to use it. Note that protobuf didn't have this issue.
This is definitely not the case. As an example, the Official Standard way to encode data structures in the 90s was ASN.1 from the ITU. Problem is, ASN.1 kinda sucked and so nobody uses it anymore outside of cryptography. Then the Official Standard became XML. Eventually people decided that also wasn't that great and switched to JSON, which is a totally unofficial standard defined by one guy.
Your approach / the EU approach would mean REST APIs would all need to be encoded with ASN.1 or XML.