Re: [Json] Duplicate names: are they erroneous or not?

[Bjoern Hoehrmann <[e-mail]>]

* John Cowan wrote:
>What I wish to avoid is a specification that says that having two
>fields of the same name in an object is a perfectly okay thing to do.
>In other words, such documents for my purposes have to be erroneous,
>rather than a legitimate way of encoding multimaps.  Is there anyone
>who can't live with that restriction?

The current text is "The names within an object SHOULD be unique". That
makes it clear that you might encounter duplicate names in the wild and
that your parser might not report it when it finds repeated names. That
is the running code situation and so I would oppose making it a "MUST".

>If there isn't anyone, then we can talk about whether this is an error
>that generators and/or parsers MUST or SHOULD detect.  But if we can't
>even get consensus on that much, it's futile to discuss parser or
>generator behavior.

People who would like JSON parsers to detect repeated names in objects
are welcome to implement and deploy and evangelise them; and when other
parsers have become the exception, then it might be useful to capture
that reality in an updated specification. Right now parsers usually do
not detect repeated names and so it would be incorrect to tell people
they can expect parsers to detect repeated names.
-- 
Björn Höhrmann · mailto:[e-mail] · http://bjoern.hoehrmann.de
Am Badedeich 7 · Telefon: [phone] · http://www.bjoernsworld.de
25899 Dagebüll · PGP Pub. KeyID: 0xA4357E78 · http://www.websitedev.de/