Hi, at some point we need to answer the questions below:
1) How many diff formats do we want inside Jackrabbit? What can we do to converge on a single one?
2) Should the diff format we use be applicable to things outside the Jackrabbit world, that is manipulating pure JSON?
Regarding 1) - as the feature set seems to be a moving target right now, this may be hard. On the other hand, if we reach consensus on certain aspects (such as: is the path quoted) we should try to update the implementations accordingly.
Regarding 2) - if this is our goal than we need to define operations in terms of JSON, not JCR, and treat the JCR-JSON mapping as orthogonal. For instance that means that there should be no special-casing of certain constructs, such as the "set" operation only taking strings or arrays.
Best regards, Julian
