I’ve been testing jsonUtility for serialization/deserialization.
I was originally using xml, with a test xml file that’s 38.1kb.
I converted that to a “pretty” json file that’s 33.4kb.
I deserialized the json file and serialized using prettyprint. The resulting file size is 3 times the size at 89.8kb because it’s using spaces instead of tabs. By replacing each group of 4 spaces with tabs, the file became 39.9kb.
I know it’s not recommended for production (serializing without prettyprint is 18.6kb, so far better) but it can still be lots more efficient if prettyprint uses tabs instead of spaces…