Re: [Json] JSON for Internet messages

[Tatu Saloranta <[e-mail]>]

On Wed, Jul 3, 2013 at 1:15 PM, John Cowan <[e-mail]> wrote:

> Tatu Saloranta scripsit:
>
> > It has never been a problem for me, nor has there been user requests
> beyond
> > schema (and schema validator) implementors who care more. But it is
> > possible that perhaps this has been a significant problem for others.
>
> The trouble is that with sufficiently bad implementations, or chains
> of implementations, such permissiveness can lead to a security breach,
> and security being a Yang Worship Word these days, everyone cowers in
> fear which such things are mentioned.  My most recent use case for
> JSON (to represent the MicroXML data model) doesn't even require parsing
> it,
> but I never suppose that for that reason JSON might as well be limited
> to nothing but arrays in which the first element is always a string and
> the second element is always an object with string values.
>


Ok. This makes more sense to me. I hope that security aspect can be
emphasized in context of duplicates, if it is the leading reason (which is
consistent with earlier discussions I think).
I also do not think there is anything wrong in having multiple documents
that add layers of requirements: strict handling of duplicates seems best
handled at higher level, and not at basic format level.

-+ Tatu +-