If you find yourself faced with hundreds of XML Schemas in your organization, you need an architecture that can help you overcome the challenges of redundant and overlapping schemas. More importantly, you need to be able to identify and reuse existing schemas rather than write new ones. Ideally, you should be able to discover and use enterprise or industry-standard schemas when available. Where you must nonetheless create a new schema, you want to be able to follow standards for terminology and structure.

This complex situation is common, as large numbers of schemas are developed for messaging middleware, Web Service remote procedure calls, application persistence, and other purposes. Often, developers in different departments write schemas without shared standards. Even where enterprise-wide standards exist, distinct companies often develop their own standards. Subsequently, they must interoperate for circumstances such as B2B integration or after a corporate mergers. and then later must interoperate for B2B integration or after a corporate merger. Worse, automated tools generate schemas with such ease that hundreds of schemas are often produced for various business purposes, without proper consideration of the need for standards.

One solution to this is using semantics for information management. This approach is centered on an information model, built using ontological standards such as OWL (Web Ontology Language). This information model is a structured representation of the business domain. Enterprise assets such as schemas -- whether XML, relational or even COBOL flat files -- are mapped to this model as a way of representing their semantics.

At this point, the semantic architecture can solve many of the asset-management problems. A search algorithm begins with business concepts in the central information model and discovers all schema elements and attributes which represent these concepts. When there are multiple schemas for a business concept, the semantic approach merges or eliminates redundancies. If that is not possible, semantically-based transformation-generation algorithms create XSLT to seamlessly integrate otherwise incompatible XML documents.

To aid in developing schemas for a new business need, the system identifies schemas for reuse, finding standard schemas wherever possible. When no existing schema is available, the system generates new schemas whose terminology and structure conform to enterprise standards.

In this presentation, participants will learn how to manage information with a semantic architecture in which all enterprise assets are unified. Although the focus will be on W3C XML Schemas, the presentation will also discuss the application of these semantic techniques to other XML schema languages, such as DTD, as well as non-XML assets, such as relational databases. The presentation will include examples and real-life enterprise case studies to bring these techniques to life.