Re: [OAUTH-WG] Possible alternative resolution to issue 26

[Julian Reschke <[e-mail]>]

On 2011-10-04 03:55, Mike Jones wrote:
> As editor, based upon James’ input, I’d like to expand the set of
> choices for the working group to consider by adding the possibility of
> using JSON string encodings for scope and error_description where the
> characters used for the encoding are restricted to the set of 7-bit
> ASCII characters compatible with the HTTPbis and RFC 2617 parameter
> syntaxes.
>
> 1. Using RFC 5987 encoding for the scope parameter.
>
> 2. Continuing to specify no non-ASCII encoding for scope parameter values.
>
> 3. Using JSON string encoding for the scope parameter.
>
> A. Using RFC 5987 encoding for the error_description parameter.
>
> B. Continuing to specify UTF-8 encoding for the error_description
> parameter.
>
> C. Using JSON string encoding for the error_description parameter.
>
> As an individual, I’m sympathetic to the argument that RFC 5987 (with
> “scope*” and language tags etc.) is overkill for OAuth implementations,
> where neither of the sets of strings is intended to be presented to
> end-users. Hence, the possible attractiveness of options 3 and C.
>
> Thoughts from others?
> ...

How is RFC 5987 encoding overkill, and JSON is not?

Note that if you'd use the \u notation inside a quoted string, you need 
to escape the backslashes due to the quoted-string syntax; which is 
likely a new source of errors.

Best regards, Julian