Re: [weirds] Ted Lemon's Discuss on draft-ietf-weirds-rdap-query-16: (with DISCUSS and COMMENT)

[Pete Resnick <[e-mail]>]

On just point 2:

On 10/29/14 11:47 AM, Ted Lemon wrote:

> Point 2:
>
> In 3.1.3, the instructions as given don't really guide implementations
> toward interoperability.   Why not just say that servers SHOULD either
> convert a-labels to u-labels, or u-labels to a-labels, and be done with
> it?
>    

See the discussion in section 6.1, referred to in 3.1.3. Converting to 
u-labels or a-labels can change the search semantics because of 
case-mapping. Remember, the server is trying to do its best in the face 
of clients that might be handing it some pretty dicey stuff. A client 
that does the right thing and only queries all-a-label or all-u-label 
strings is going to get consistent results, because it is taking 
responsibility to do the appropriate mappings. But if the client is 
blindly handing over labels it got from somewhere else (which is why you 
would get mixed stuff), the server is going to do what it can to get 
things right (or decide that it doesn't want to try). So the most 
interoperable thing is to say, "Clients: Query reasonable things. 
Servers: You can choose to be helpful or not on unreasonable things." At 
least that's my understanding of the consensus of the WG and the IDN 
experts consulted; the IDN experts on the list can pipe up if I've 
gotten this wrong.

pr

-- 
Pete Resnick<http://www.qualcomm.com/~presnick/>
Qualcomm Technologies, Inc. - [phone]