An Integrating Architecture for Competence Management

Introduction

Software applications dealing with human resources and their skills, attitudes, and knowledge (such as e-learning systems, skills databases, e-recruitment portals, corporate portals integrated with competence-centered services, and functionalities) are often based on database technology (usually relational) for storing, organizing, and searching relevant information. These existing applications have two major limitations. First, they and their databases are based on raw data (such as CVs and job offers and job/role descriptions), which are organized according to some ad-hoc “reference grid” (like a job or trade tree): indeed, limited attention is devoted to data organization and to its foundations. Data organization should be based on the central concept of competency: raw data are interesting if they convey information about what abilities are required for accomplishing tasks and what abilities individuals hold (or have acquired); this information is indeed forming the competence, required and acquired respectively. Second, applications based on database technology do not really support the systematic analysis, exploration, and sharing of raw data and therefore offer limited support and weak integration to what can be called competence management processes. For instance, within a process for assessing individual competencies, it is difficult to implement portal services that try to automatically find out competencies of individuals from their CVs or, inside a company, from other documents (like activity or process reports which individuals have made).

Unfortunately, despite a huge amount of work, there is no consensus on the competency definition. This is especially because most of the current work prioritizes some processes over other processes (e.g., evaluating competencies is prioritized over identifying needed competencies). Worse, as usual, some works prioritize enabling technologies over models. This results in partial or overloaded models for representing what competency, acquired and required, is; additionally, there is no clear picture of limitations of these models because prioritized processes are often hidden.

For these reasons, we have developed since 1999 (Harzallah & Vernadat, 2002; Harzallah, Berio, & Vernadat, 2006), a competence management process reference model (Harzallah & Berio, 2005) and a competence reference model (the CRAI model, competency, resource, aspect, individual). These two reference models:

• •

  Clarify what competence management should provide, therefore providing a management model to be put in place by organizations willing to manage their activities and to make decisions also in term of individual competencies required and acquired;

• •

  Provide technology-independent precise definitions for concepts of individual competency both acquired and required; however, these definitions can be operationalized by using several technologies;

• •

  Are complemented with guidelines for deploying them in one organization.

The article describes and discusses an integrating architecture for competence management—introduced earlier (Berio & Harzallah, 2005, 2007)—originated from the two reference models cited. While the reference models allow to identify (or to assess) which services and functionalities should be implemented—now and in the future—in human resource software applications and how these services and functionalities interact, the integrating architecture represents the starting point for incremental and iterative implementation—driven by specific organization priorities—of those services and functionalities. Additionally, the integrating architecture takes into account, advocated or reported in literature, design principles, and the recent techniques and standards for implementing these services and functionalities.

The article also provides a concrete example of how this architecture can be used for justifying and adopting one specific technique, that is, dynamic taxonomies, to implement identified functionalities.