Re: [ietf-types] Registration Of media types application/cdmi-*

[Keith Moore <[e-mail]>]

On Sep 26, 2010, at 9:03 PM, Krishna Sankar (ksankar) wrote:

> Keith,
> 	Thanks. Good comments. Corrected.
> 

Thanks, but I have some additional comments.  Taking just one of these as an example:

> 6.1.  Media Type application/cdmi-capability
> 
>   Type name: application
> 
>   Subtype name: cdmi-capability
> 
>   Required parameters: none
                        ****
>                      
>   Optional parameters: none
                         ****
> 
>   Encoding considerations: The "charset" MIME parameter, if present,
>   MUST be set to "UTF-8", as defined in [RFC3629].  

You've declared that there are no parameters, either required or optional, and then you immediately say that the (nonexistent) charset parameter must be set to UTF-8.

Normally application/* types don't have a charset parameter at all.  charset is a parameter that tends to be used with text/* types.  RFC 3629 even says (in section 8) "UTF-8 is suitable for use in MIME content types under the "text" top-level type."

You can define a charset parameter for your type if you want to do so.  But if the charset is always going to be UTF-8, I don't see why you should bother.  Putting a charset parameter there just invites people to set it to something other than UTF-8.

Also, if you want to insist that the representation always be in UTF-8, the best way to do that would appear to be to simply state that the representation is always UTF-8 and that any charset parameter is to be ignored.