تعبیر معانی در اتوماسیون مدیریت منابع انسانی از طریق SQL Embedding semantics in human resources management automation via SQL

1 Introduction The so-called network economy changed the way of conceiving knowledge management in organizations. Enterprise borders go far beyond its physical location and more and more knowledge-intensive resources become available across the network. Such a shared environment asks for unambiguous interpretation of information sources and makes knowledge-based technologies a key success factor. In particular, business processes automation can take a crucial chance by technologies supporting knowledge representation and management [20]. Among other business activities, human resources management can get a significant boost from the adoption of knowledge-based technologies aimed at business automation. In fact, intellectual capital is an asset with peculiarities: it is intangible, its description is subjective, and the way to describe such an asset is a key choice for its successful exploitation. Logic-based representation perfectly fits such peculiarities, and the representation language—if properly chosen—could support human resources management automation through suitable reasoning services. Even though such flexibility and informative potential is paid in terms of computational cost of reasoning, semanticbased enterprise systems may represent a good candidate to substitute traditionally employed solutions, mostly based on Relational Data Base Management Systems (RDBMSs). Despite recent emerging systems in the data storage field, such as NOSQL1 technologies, we focus on RDBMS, as structured data model engine, and SQL, as declarative language for data manipulation and retrieval, since they still represent the most popular and well-known tools in enterprise scenario. Nevertheless, RDBMSs lack of a semantic characterization of resource descriptions thus allow only for syntax-based information management, which is too restrictive for a domain as knowledge-intensive as human resources. Our main goal in this paper is aiming at filling such a gap. For example, available e-recruitment tools2 generally store information about employment, personal data, certifications and competence of candidates by exploiting RDBMSs with customized and structured templates. Then, some information is extracted through relational query languages—SQL or some variants. Now, an RDBMS is surely suitable for efficient storage and retrieval of data, yet SQL is usually not flexible enough to support a discovery process as complex as recruitment. Take, for instance, the problem of selecting the most appropriate candidate for some job; this is a process involving several preferences, some along orthogonal dimensions—e.g., acquired skills vs. geographical location—while others somehow related— e.g., learning gaps and expected salary. By means of SQL standard operators such as aggregate functions, group by and order by clauses, the user is able to retrieve the best tuples (i.e. the best candidates for a specific task) according to her sorting criteria; yet such tuples still need a human-based post-processing phase, where incompatible or unsatisfied preferences (not directly representable by standard SQL) are evaluated and traded off.