Can you offer some detail into why you find this approach useful? From an outsider's perspective, if you're sending around JSON documents so big that it takes

Can you offer some detail into why you find this approach useful? From an outsider's perspective, if you're sending around JSON documents so big that it takes so long to parse them to the point reordering the content has any measurable impact on performance, this sounds an awful lot like you are batching too much data when you should be progressively fetching child resources in separate requests, or even implementing some sort of pagination.