Re: implementations of "XML Voucher: Generic Voucher Language" ?

[Ko Fujimura <fujimura@isl.ntt.co.jp>]

Hi Lynn,

Thank you very much for providing us a lot of X9.59 information.

I understand that the encoding rules of the spec will be helpful if XML Voucher 
have to define XML-encoding rules of a signed object...

But, the digital signature (or signed object) is used only in the delivery of 
Voucher Component or within the VTS-specific protocols that rewrite the ownership
of a voucher in the current XML Voucher processing model. In either cases, the 
XML-encoding rules of a signed object is outside scope of the XML Voucher 
specification (e.g., Encoding rules defined in XMLDSIG is another candidate).
This language design is one of major issues to be discussed in the next IETF 
meeting, though.

Thanks,

Ko
lynn.wheeler@firstdata.com wrote:
> 
> there were/are two parts of FSML ....
> 
> 1) deterministic encoding rules for tagged grammers so that the signed object
> and the object having signature being verified will work in circumstances where
> the intermediate message have have totally mangled the original signed object.
> 
> 2) echeck
> 
> on the other hand, X9.59 standard was targeted at all account-based electronic
> payments; credit, debit, ATM, echeck, ACH, wire-transfers, internet,
> point-of-sale, wireless, etc.
> 
> The X9.59 standard calls for ASN.1 encoding.
> 
> A mapping of X9.59 to ISO8583 shows the original object being signed and then
> the fields from the original signed object being transmitted as ISO8583 fields
> (not in the format of the originally signed object) and then the originally
> signed object is reconstructed at the relying party and the signature verified.
> 
> An X9.59 object using  XML-ecoding rules could be implemented. However, in the
> case of the mapping example of X9.59 to ISO8583, the reconstruction of the
> signed object at the relying party has to be deterministic. As far as I know,
> FSML was the original tagged definition that specified deterministic encoding
> rules ... that would allow tagged-encoded objects to be reconstructed in a
> deterministic manner (therefor the FSML reference; not because of the echeck
> specification but because of its specifying deterministic encoding rules for
> tagged fields).