It seems my post is not being accepted by the OASIS Mailer, probably because I'm an observer and not a member. Perhaps one of you could post it on my behalf.
. . . . . . . Ken
Date: Tue, 10 Mar 2009 07:28:34 -0400
Subject: Re: [ubl-psc] Requirement for Document Reference/External Reference ABIE
At 2009-03-10 12:08 +0100, Oriol Bausą wrote:
In this ABIE, there is a DocumentHash element that has DocumentHashType data type. This data type is an extension from Text.type. The point is that is required being able to define the algorithm used to generate the hash from the document.
Nevertheless, I see that this would break the CCTS architecture, so we should define it using another BBIE, something like:
<cac:DocumentHash>
<cbc:DigestValue>qZk+NkcGgWq6PiVxeFDCbJzQ2J0=</cbc:DigestValue>
</cbc:DocumentHash>
May I offer that since we already have cbc:DocumentHash, then we only need a sibling cbc:HashSchemeID optional BBIE? I agree the pairing of the two items implies the use of an ABIE if we were designing a new model. But for backward compatibility and preventing the duplication of the same semantic in two entities in UBL, I don't think we have the flexibility to introduce an ABIE in place of or in addition to the BBIE.
I only see cbc:DocumentHash in cac:ExternalReference, and in there we already have sibling items ExpiryDate/ExpiryTime rather than an ABIE for Expiry, so I think the following in 2.1 wouldn't make users uncomfortable:
cac:ExternalReference
cbc:URI
cbc:DocumentHash
cbc:HashSchemeID External Reference. Hash Scheme Identifier. Identifier
cbc:ExpiryDate
cbc:ExpiryTime
While I prefer the word "algorithm" when I write prose, I think a precedent has been made in CCTS to call such things schemes.
I hope this helps.
. . . . . . . . . . . Ken
G. Ken Holman
Crane Softwrights Ltd.