Document Type : Original Article


Assistant Professor, Knowledge and Information Science, Faculty of Psychology and Educational Sciences, Allameh Tabataba’i University, Tehran, Iran.


Purpose: Library users need to have integrated access to digital library content and services. Preparing these services needs syntactic and semantic interoperability infrastructure. Nowadays most digital library applications syntactically have the ability to exchange with each other. Therefore the purpose of this paper is to propose semantic integration architecture for digital libraries to enable them to prepare new semantic services for their users.
Method: This study was carried out in the following steps: firstly research literature was collected and studied. Studying research literature helps the researcher to prepare a primary version of the checklist to collect data in this study. The next step aimed to survey the semantic interoperability ability of digital library management systems. In this regard, 26 Iranian digital library management systems were studied by the researcher-made checklist in the previous step. The studied digital library management systems have syntactic interoperability functionalities. Then this study’s semantic integration architecture was proposed according to the analysis and results of the study. Finally, the "demonstration" method was used to assess the proposed architecture.
Findings: Findings showed that studied digital library management systems focused on textual materials and paid very little attention to other library information resource formats. Moreover, in many cases, the studied digital library systems did not fill the contents of semantically needed metadata fields. On average, they attempt to document metadata fields such as subject and keywords, by existing reference materials such as Library of Congress Subject Headings and subjective thesauruses that they needed.
Conclusion: According to the findings, the architecture of the semantic integration model based on the results of this study was proposed with three data, inference machine, and application layers.