We at IVANS were fascinated with the lively debate on the ACORD XML standards and their development between Kevin S. Kelly and Josh Lee of Microsoft in the June issue of Technology Decisions, and Gregory Maciag and Ron Dudley of ACORD in the July issue. As active participants in ACORD, we would like to contribute to this debate.
If anyone thinks, as Microsoft apparently does, that all or most of the value of the ACORD XML standards is in the data dictionary, then that dictionary is certainly there and available for use. It has been officially there for a year now. In P&C, that is where ACORD started, and that was the first thing finished.
However, having a data dictionary without predictability as to the message structure doesnt give assistance to receivers of standard messages. Its easier to deal with varying tag names and a predictable structure than predefined tag names and an unpredictable structure. The former is a straightforward one-to-one translation; the latter requires semantic leaps and assumptions that lead to confusion when different people process the same message differently. For example, if you call a sublocation a building, but you know that the terms have the same meaning, thats easy to handle. But if you neglect to specify where a vehicle is located or who is the primary driver on a vehicle, or if you convey this information in different ways (Sender A uses references and Sender B uses containment, for example), youve defeated effective data communication.
The worldwide Internet standards that the Microsoft article holds up as models have the same sequencing requirements and hierarchy restrictions that the authors criticize in the ACORD standards. The XML Schema Standard took three years to develop, and it lays down a structured data model by which to define schemas. SOAP and WSDL prescribe very hierarchical structurefar more than just a data dictionary. XSLT has strict rules about which attributes apply to which elements, and which elements can be children of which parents.
These global XML standards also take time to develop and implement. Version 1.1 of the SOAP standard was proposed to the W3C in May 2000; Version 1.2 is not quite official as of this writing. Almost all vendors SOAP Toolkits (even Microsofts) still implement Version 1.1. So ACORD and the insurance industry arent doing badly by comparison.
The Microsoft article says data modeling should be left to insurance software developers. The ACORD standard, through its SPX facility for defining company-unique data fields and edits, gives developers great flexibility. It does so in a way that allows receivers of the data to process the data without out-of-band communication, since it allows metadata (data types, lengths, usages, hierarchy position, etc.) to be specified in-band using a standardized XML syntax.
The Microsoft article advocates that simple transactions such as address changes should be possible without sending a complete policy image. This was tried in the AL3 standard, and it was found that nearly all carriers needed a full policy image to process the requested change. Part of the problem is that few changes in insurance are really localizedif you change an address, youre probably changing the location of a number of items at risk (vehicles, jewelry, etc.), which changes rating factors and premiums. If you add a driver to a policy, youre changing the driver usages for every vehicle on the policy, not only for the new driver but for all the old ones. Since agency systems need to store an up-to-date image of the policy to make policy change processing possible at all, its not a major problem to provide a complete image. The standard defines structures that allow receiving systems to make easy identification of the specific changes in the stream.
With regard to the standard-setting process, which the Microsoft article also criticizes, we can attest as non-ACORD-staff participants that volunteers from the industry have had a predominant role in the formulation of the entire P&C XML standard, and the assistance of the ACORD staff has been invaluable.
We think the ACORD standards, at least for property/casualty, are on track and will be used as extensively as AL3 has been. In no small part, this is because of Microsofts participation, and even more because of the strength of Microsofts XML software tools and their faithfulness to cross-industry standards.
We welcome contributions to Trade Talk. Articles should not exceed 750 words, deal with issues only (no product pitches, please), and relate to technology as it affects the insurance industry. E-mail contributions to tradetalk@tdmag.com. You will receive a reply only if your submission is chosen for publication. The editors reserve the right to edit for space and clarity.
© Arc, All Rights Reserved. Request academic re-use from www.copyright.com. All other uses, submit a request to TMSalesOperations@arc-network.com. For more information visit Asset & Logo Licensing.