The Advantage of Openness
How Open Semantic Platforms Improve Business Performance.

There are parallels to be drawn between the way we manage unstructured information today and the closed data world of the 1970s, before relational databases and business intelligence layers. In the ‘70s, data was unlocked by programmers and the business had to wait every time something new was required.

Today unstructured information has accumulated at an incredible rate, often unmanaged and held in multiple repositories. Again, specialized agents with a hard-earned understanding of the systems below are needed to unlock their secrets.

Business is again reliant on specialized skills in the hands of a few to unlock their information assets. As before, an opening up is required to properly harness the value of organizations’ information assets.

Just as the gap between closed and open data was filled by new approaches in the ‘70s and ‘80s, so too will the gap between limited search and productive find.

This new approach requires semantic middleware.

Searching Without Understanding
Today’s environment consists of two distinct layers; the human interface and the content services. Lots of users, from multiple disciplines, interact with different interfaces offering alternative methods of access to information.

As content is created, its metadata (the labels that describe the information) is manually applied by editors and authors. Often the metadata is non-existent, or inconsistent across systems or inaccurate.

Given this environment, searching for information can be time-consuming, frustrating and ultimately dissatisfying. After searching, the pertinent information may be missing from the result or lost among hundreds of results.

People use language descriptively, idiomatically, ambiguously and tainted with jargon. Search engines are binary and lack any idea of a subject’s meaning or the way humans use language. Search takes the few words that are offered at search time and, using closed algorithms, scans the index for these words—literally.

Is “orange” a fruit, a color, a brand name or a place? Is “wound” an injury or balling up some string? The existing search experience is lacking context, and context is key to finding. Adding context to the mix improves information management and “findability” measurably.

But defining the context—the semantics—is complex, because language is complex and applying it takes effort. To be commercially viable, tools need to be employed to automate and assist.

The Case for Semantics
Most organizations recognize the need to establish classification standards so that standard metadata values are applied to content regardless of its origin.

Automating the application of metadata ensures consistency. Consistent metadata allows information stored in one repository to be joined with similar information in another. Consistent metadata means the user can be confident that all the information on a subject is presented.

There is a case for opening up the metadata for use by other systems. For example. it might suit that a document needs its metadata stored with it, as a matter of record. There is also a case to enhance the user’s search experience: suggesting related topics, filtering to truly relevant subsets and providing a taxonomy path so that the user is kept in context with the subject that they are researching.

There is a case for a new layer that lies between the user interface and the content technology.

This new layer is the open semantic platform. “Open,” as it needs to offer its facilities to any external system, “semantic” because it is about adding meaning to the information and “platform,” because it lies between the human interface and the content services as a new layer—available to any application that needs it.

An open semantic platform is an ontology management system that enables the enterprise to maintain controlled and social vocabularies (a.k.a. semantic models), that describe the information domain, provide the context and form the standards for information classification.

It is a classification system that takes the semantic models and uses them as the evidence base for information classification. This classification approach should be transparent, auditable, adjustable and ultimately accurate—enhancing the metadata quality. Web service interfaces to the classification engine should ensure the metadata is available for use by the search engines and/or the content management systems and/or assisting workflow processing.

Natural language processing tools and text mining capabilities increase the productivity of the taxonomy and classification development to make the process commercially viable.

To offer a contextual navigation experience, the model also needs to offer its terms and structures through open APIs to user portals, search interfaces and line-of-business applications. By ensuring that the same model that is used for classification is also used for navigation, the system achieves a strong positive feedback loop that ensures context and relevancy that delivers an exceptional user experience and findability.

The open semantic platform delivers advantage by making it possible for a business to model its domain, and then organize, automate and communicate its information according to this model. This means that search engine precision is enhanced, the quality of metadata is improved, CMS implementations are sped up, unstructured information can be integrated from multiple repositories, compliance processes can be automated and the user experience that is offered to clients, staff and partners will be exceptional.