To:
janusz sienkiewicz <janusz@ca.afilias.info>
Cc:
IETF Provreg <ietf-provreg@cafax.se>
From:
Joe Abley <jabley@isc.org>
Date:
Tue, 6 Apr 2004 17:07:47 -0400
In-Reply-To:
<407319E4.A028F961@ca.afilias.info>
Sender:
owner-ietf-provreg@cafax.se
Subject:
Re: [ietf-provreg] Unique identifiers for Contact
On 6 Apr 2004, at 16:58, janusz sienkiewicz wrote: > Joe Abley wrote: > >> Are there server implementations in the wild which return errors if >> the >> client-supplied contact:id element is empty, or contains a >> non-server-unique string? >> > > I think an EPP standard compliant server implementation should return > error > response if the client-supplied contact:id element is empty. The XML > type > for the element is clIDType and the type has minimum length which is 3. Ah, good point. > Assigning a value by the server in case of non-server-unique string > could be > very confusing for some clients. It is probably safer to return error > code. The spec seemed fairly clear to me that clients shouldn't expect the server-unique string in the response to be the same as the server-unique string they suggested, but upon re-reading, it does seem reasonable to not make that assumption. Scott: what did you intend, here? Janusz: I like your proposal to make <contact:id> optional, too. I wish someone had suggested it before last call :-) Joe