1 Executive summary

This discussion paper outlines the complex-interrelated issues around Controlled Value Lists (CVLs) and their use in eXtensible Markup Language (XML) programming. This paper is designed to offer a basis for wide discussion onhow to use and maintain CVLs and how to leverage their creation and re-use in XML schemas. Common agreement on the process will reduce development time and improve the quality of data exchange across government agencies.

CVLs in records management versus XML programming

CVLs are used extensively in records management, as the basis for metadata creation, where they are commonly referred to as Controlled "Vocabulary" Lists. (In XML programming, they are also sometimes called Controlled "Code" Lists.) The needs of records managers are similar to those of XML programmers, although the outputs are different. Records managers define and reuse terms to create metadata, which improves search and data control. XML programmers define terms to create reusable schema fragments that can be referenced from other schemas. The more programmers reuse the same schemas to describe the same elements, the better the quality of the data exchange.

In the records-management environment, the Thesaurus Advisory Group (TAG) already exists as an expert CVL management and maintenance forum for CVLs in the New Zealand Government Locator Service (NZGLS). There is no equivalent for XML schemas. In XML-schema programming within government agencies, the lack of a clear process to agree, access, update, and maintain a common set of CVLs is one of the hindrances to successfully implementing reusable schema fragments.

CVLs in government need common terms

There are dozens of CVLs and data dictionaries being developed or in use today across government, primarily for data-administration purposes. Examples include Statistics and Education; other sectors such as Health are commencing work on them. It is much easier for sectors to agree upon their specific terms than it is to agree upon a generic term, such as a street suffix. Unfortunately, the generic terms have wider application in XML schema development.

The issues in CVL standardisation are significant, not least of which is agreeing upon which terms are to be in the list and which terms are excluded—before even trying to identify and make accessible all the CVLs and data dictionaries for XML schema creation. Superficial investigation indicates there are differences between terms used in vocabulary lists and those needed by XML programmers; for example, there is no common list of street suffixes such as Street, Lane, Crescent, Place, etc. in NZGLS. This example is demonstrated in 3.2

Towards forming a Working Group and Terms of Reference

This paper calls for a preliminary discussion and lively interchange of ideas on CVL management for programmers. It also seeks to locate specialists in this area—wide circulation in both the public and private sectors is encouraged. Outputs from this process will be used to establish a Working Group and to formulate its Terms of Reference. The Working Group should include records managers, data administrators, solution architects, and XML developers working in or for the public sector. The requirement is to get a common understanding on the extent and role of CVLs—and how to access, exploit, and maintain them.