Re: [domainrep] I-D Action: draft-ietf-repute-media-type-13.txt

[Barry Leiba <[e-mail]>]

>
> There's one thing that should be brought to the WG's attention, in case it
> will be a problem for anyone: The JSON format described in the media-type
> document changed a bit on the advice of a couple of IESG members more
> well-versed in such things than I am.  Anyone interested should go have a
> look and say something now if it's wrong-headed or going to create problems.
>

In particular, since I was one of the ADs who asked for the changes, I'd
like to say a few words about them.  There are three main points:

1. The content-type parameter was moved into the JSON object.  It seemed
wrong to overload the media type wrapper with something that really belongs
in the payload.

2. The set of reputons in the response was changed from a bunch of JSON
objects that all have the same name to a JSON array.  JSON says that object
names SHOULD be unique, and it seemed wrong to violate that.  And as long
as the object names don't matter, it makes more sense to use an array.

3. The mechanism for defining the JSON format is no longer ABNF.  The JSON
spec already defines the ABNF for JSON, including all the delimiter
characters, the DQUOTEs, where the whitespace can go, and so on with other
fidgety bits.  It seemed wrong and risky to re-define all that here.
 Instead, the JSON format is now defined in terms of JSON primitives.

I think these are all important changes that go in the right direction.
 But the working group should review that, and anyone who disagrees should
raise an issue.

Barry, Applications AD