I really wish things like this wouldn’t call themselves “a superset of X”.
It’s a different data format, that happens to overlap JSON, and while technically that makes it a superset, that’s not what the actual result is.
For example: is calling a plain text file a “superset of xml” a useful claim to make?
I’m also not sure I understand the purpose of this or what it is trying to solve - everything it has clonked on to “Json” already exists in a much more clearly standardized way in xml. XML also has numerous production quality parsers, and a huge ecosystem. The only thing it’s missing is a compact binary form - that would be a reasonable argument for a new structured data format. But instead what they’ve done is published a new “plain text Json superset” that /just happens/ to have a binary serialisation.
"I’m also not sure I understand the purpose of this or what it is trying to solve"
From the 'why' page, it sounds like a middle ground between JSON (max flexibility) and protocol buffers (max efficiency of network bandwidth and parsing workload):
http://amzn.github.io/ion-docs/guides/why.html
It’s a different data format, that happens to overlap JSON, and while technically that makes it a superset, that’s not what the actual result is.
For example: is calling a plain text file a “superset of xml” a useful claim to make?
I’m also not sure I understand the purpose of this or what it is trying to solve - everything it has clonked on to “Json” already exists in a much more clearly standardized way in xml. XML also has numerous production quality parsers, and a huge ecosystem. The only thing it’s missing is a compact binary form - that would be a reasonable argument for a new structured data format. But instead what they’ve done is published a new “plain text Json superset” that /just happens/ to have a binary serialisation.