The subjects handled within an enterprise are listed through classificatory schemes, document indexes, thesauri, classifications of document centers, Lotus Notes databases, keywords in document management systems, domain-specific XML standards, etc.

The core component of any document organization scheme, the subject offers a reference language within the enterprise. The exact definition of the subject, managing its expression in different languages, using different words to talk about it depending on the departments, are problems that are never wholly resolved but which demonstrate the importance of the subject and its definition.

The subject may be a general concept, a technique, a product, the name of a process, a profession, a person, a project, a company that is a client or supplier. The totality of subjects may be divided among:

The subject is a basic tool for:

Just as in the case of ordered and systematic access to books in a library or bookshop, content organization relies on the organization of subjects in the enterprise. This organization has existed since a long time in the enterprise and is found scattered across different systems of document management:

These different resources for organizing subjects use differing methods:

Part and parcel of the culture of the enterprise and its employees, these different elements of organizing subjects must be federated and incorporated within the Content Intelligence system of which they form the infrastructure

We are considering here what happens within a single enterprise, but the enterprise is global, composed of several entities, departments. The content management systems are numerous, dispersed and specialized:

The aggregation of content is necessary in order to meet the content consumer’s need to access all the relevant documents throughout the enterprise.

For the responses to precise subjects to be pertinent and exhaustive, there must be:

Building an enterprise-wide subject repository is first of all a work of reusing whatever subjects already exist:

Automatic taxonomy creation tools that mine a corpus of documents may help enrich and control the list of subjects.

Each subject must be defined, along with its different synonyms and orthographies, so as to take into account the different languages used in the enterprise.

The management of subjects requires software tools that are adapted for administering these subjects, for updating and deleting them, and so on. The software solutions for managing subjects must include functions for collaborative work, workflow, and the automatic importation of legacy data. They must likewise be sufficiently standardized and open so as to be used both by humans and through automated handling by other software components of a Content Intelligence solution.

Subject organization relies on different tools depending on the nature of the subjects to be organized:

The creation of one or several systems of subject organization may thus rely on the reuse of elements already existing in the enterprise.

Managing the organization of subjects requires software tools adapted for this specific purpose. These tools must be capable of administering hundreds of thousands of relations that organize subjects, of reusing pre-existing organizations that derive from ontologies and thesauri, of mapping the relationships described in the databases.

These software solutions must be open enough to serve as the basis for building tools that handle requests and for transacting with the various other components of the content intelligence solution: natural language searches, inference engine, tool for graphically representing the organization of content, etc.

Software solutions of the inference engine type may likewise be implemented so as to enrich the organization of subjects from already existing organizational elements.

The referencing of content relies on different techniques depending on the nature of the subject:

The solutions will be different depending on the nature of the documents:

For all such contents referenced in a stable manner to subjects, an adequate software solution should allow the enterprise to manage these relations between the subjects and the contents, and to place these relations at the disposal of the other components of the Content Intelligence solution.

Document search using search functions should combine:

Natural language search, based on a defined professional vocabulary and against subjects and their organization, seems to us an indispensable tool as soon as the end-user attempts to search within domains with which she is the least familiar.

Navigation functions between the subjects (broader or narrower subjects, different parts of a machine, patents linked to a molecule, successive steps of a process, technologies used for a project, etc.) are of the utmost importance for finding the correct information when the search has made it possible to identify the subjects.

These navigation functions may be wholly manual by using the hyperlinks on the navigation pages, but may also be automated.

A simple automation would allow all the documents of a certain type for a product to be found along with its hierarchically linked components. By relying on navigation automations or on inference engines, a more complex automation of the navigation would allow all semantic and logical relations between the subjects to be explored.

Components for graphically representing subject organization should likewise be included within a Content Intelligence solution, whether to represent relations between the product components or a manufacturing process.

The extensions of the Topic Maps standard regarding the identification of subjects (Published Subject / OASIS technical group) allow to manage, in a centralized or decentralized manner, the stable and unique identifiers of the subjects.

This basic component is of utmost importance to ensure:

The separation between:

allows to decentralize the solutions for managing subject organization and those for managing contents, and thereby ensures their mutual independence.

One could have a general organization of the enterprise subjects so as to aggregate the content of generalist intranets and also, for example, a subject organization relating rather to the technical documentation of products The ability to manage these two systems independently, in an autonomous manner, is in fact indispensable.

Any system of referencing subjects, organizing subjects and content access should include personalization mechanisms that ensure the following functions:

The Topic Maps standard along with its derived tools allow for these personalization mechanisms.

Enterprises already have various sorts of content management solutions in place, with a multiplicity of systems, and they need to federate these diverse contents. This is why implementing solutions that would require a modification of these existing systems is out of the question.

The principles underlying the Topic Maps standard and the software that implements this standard are based on the referencing of contents through subjects, without there being any need to modify the content management.

The Topic Maps solutions moreover allow to manage a set of associated metadata for each document without modifying the content management solutions already in place.

Deriving from the world of documents and having its origin in the problems posed by the reuse of pre-existing document organizations, the Topic Maps standard does not impose any particular model of organizing documents. The association mechanisms support all possible models of knowledge organization by taking the current needs and existing state of affairs into account, and enabling the coexistence of these different models:

thus allowing just as well the description, within a single system, of the document classification, the organization of a complex thesaurus, the breakdown of a machine, the relationships between the participants in a project, etc.

We’ve seen that a Content Intelligence solution puts to work several technologies:

These tools, taken as a whole, use:

Though each of these different tools uses its own techniques and handling algorithms, a close coordination must be maintained between the tools and the Topic Maps solutions so as to ensure:

This coordination requires a normalization of transactions through standardized XML files, API standards, and Web Services.

Topic Maps tools are ideally suited for providing the infrastructure to content intelligence solutions because of:

These strengths explain the proliferation of operational projects based on the Topic Maps standard and tools within various industries, the publishing world and in administration.

Several elements are still missing in the Topic Maps standard and its derivative software for the solutions implemented to be complete and interoperable:

But, as of today, these different challenges have already been identified by the different actors and are progressively finding solutions within the framework of standardization committees, operational projects for clients and industrial collaborations among the different providers of technical solutions.

At the core of the enterprise, Content Intelligence tools will make it possible to: