Re: [Json] Questions regarding duplicate names

Agree with chairs’ outline of the rough consensus. -T

On Sat, Jul 27, 2013 at 1:22 PM, Matt Miller (mamille2)
<mamille2@cisco.com>wrote:

> Hello all,
>
> Here is an attempt to summarize responses to the duplicate names questions.
>
> The consensus from this thread seems to be that duplicate names within the
> same JSON object are not generally desirable.
>
> There appears to be rough consensus that a generator (for some definition
> of "generator") ought to take steps to avoid emitting duplicate names, but
> that there exists a non-trivial amount of software where the prohibition of
> duplicates cannot be completely enforced.
>
> There appears to be consensus that parsers (for some definition of
> "parser") need to be able to deal with duplicate names in some manner.
>  There is rough consensus that 4627bis ought to provide guidance on how
> parsers deal with duplicate names.  However, there is no one option that
> has any consensus; instead, there appears to be acquiescence that a given
> parser ought to support one option from a limited set of documented
> choices.  The most common suggestions -- which appear to reflect the
> majority of existing implementations -- far so:
>
> * report an error in some manner
> * report the last value that appears in the JSON text
> * report all the values that appear in the JSON text
>
>
> - Paul Hoffman and Matt Miller
> _______________________________________________
> json mailing list
> json@ietf.org
> https://www.ietf.org/mailman/listinfo/json
>
>