>some sort of agreement over the structure [for json] and wide support for it
there isn't though: i had to adjust the code to account for what they apparently meant as opposed to what was documented, given examples that people produce. it was similar but a bit worse for ogdl.
