On 2012-04-05 16:31, Stefan Guggisberg wrote:
> ...
Well.
If "/" is allowed in object names (in a role other than a path separator),
we need to be able to escape them.
If we don't need them, there's nothing to worry about in practice, as the
escaping will never happen for us...
/me confused ;)
i agree that in our current use of this format within oak it won't be a problem
since "/" cannot occur in a name.
however, you said that if JSOP diff (or whatever it's going to be
callled) is positioned
as a generic patch format for JSON, we do have to deal with ^/ in paths/names,
i.e. JSOP protocol implementations need to be able to handle it.
*If* it is positioned as that.
I believe that's David and a few others want, but at some point of time
we need to decide on that.
are you arguing for positioning JSOP diff as a generic patch format for JSON
or against?
I personally don't care.
I believe the reduced verbosity isn't worth defining a second format (in
addition to JSON-Patch), but if people like it so much we can do it. But:
- we need to decide, and
- if we do it, we need to bite the bullet and make it generically
applicable.
David?
Best regards, Julian
