On Fri, Oct 21, 2011 at 7:47 AM, Paul C. Bryan <paul.bryan@...
> I've posted the second draft of the JSON Pointer Internet-Draft to the
> It should address all of the outstanding issues that have been raised to
> date. Your feedback is welcome.
i don't know if this would be useful in the larger context of your RFC, but
rather than reserving a specific path character (since library-level code
cannot expect arbitrary JSON to conform to one), in my code where i use
"path-like strings" to search for JSON data i instead require that the
caller specify the separator character (to avoid awkward encoding issues).
So, instead of a path like "foo/bar/baz", i require "XfooXbarXbaz", where X
is any character guaranteed (by the client, not the library!) to not appear
in any key name in the path. That X is then used as the separator for the
search string. This approach does not address absolute vs. relative paths,
however, and assumes that all searches are relative to an object provided by
----- stephan beal
[Non-text portions of this message have been removed]