JSON started out with JavaScript and slowly expanded beyond the ECMAScript specifications to become more versatile. I don't see what the big deal is.
Humans prefer easily processable and storable information, and for this reason they like to think JSON is JavaScript Object Notation.
This reminds of how ECMA doesn't like to be called the _European_ Computer Manufacturer's Association, since they have now grown beyond that, and prefer being called Ecma.
As the OP says, the big deal is that people frequently assume that JavaScript's eval is an acceptable JSON parser — for example, this is the fundamental assumption which makes JSONP work. If your JSON isn't guaranteed to be valid JavaScript, you might run into actual problems.
Wouldn't using eval to parse JSON be quite hazardous in the same way as python's pickle, and thus be limited to pretty specific use-cases (limited moreso than just simply to JS)?
It used to be quite common and is still done (and is required for the hack known as JSONP); lots of the time you're retrieving JSON from the same server you loaded the original page from so it's safe.
Humans prefer easily processable and storable information, and for this reason they like to think JSON is JavaScript Object Notation.
This reminds of how ECMA doesn't like to be called the _European_ Computer Manufacturer's Association, since they have now grown beyond that, and prefer being called Ecma.