Re: [kmip] Use cases for mutating attributes

I agree with Matt, some of the use cases proposed should result in an error,
the server should not make assumptions about what the client wants/needs, the
server should tell the client that the request was invalid and make the client
adjust the request.

-Wyllys


Matt Ball wrote:
> Offhand, I think in many of these use-cases, it seems that it would be
> better to return some kind of error than to quietly 'correct' the
> attributes.  In the example with a 3DES key protecting a petabyte of
> data, it would be better for the server to return an error than to
> correct the value to what the server assumed that the client had
> originally intended.  Maybe the client really wanted an AES key, and
> accidentally chose 3DES.
> 
> As a group, we need to be very careful about what attributes we allow to
> change, because it's likely covering up some other systematic problem. 
> These are often use-cases that should result in an error.
> 
> Cheers,
> -Matt