3.3 Need for single window search interface
The repositories may be different in their coverage, software usage, nature of contents and most importantly in retrieval techniques and tools. As a result, it is difficult for end users search comprehensively these repositories that provide scholarly materials freely. This situation calls for the development of a single window search service covering all the repositories in a given domain of knowledge. Of course the repositories need to be compatible with interoperability standard(s) for building a search service on the basis of harvested metadata. These single windows search services (based on resource metadata) are advantageous to scholars and others as it brings them closer to uniform access interface for scholarly information bearing objects and cultural resources (Cole, 2003). Directory of Open Access Repositories (OpenDOAR) shows that there are around 2606 open access repositories as on 2014 and Registry of Open Access Repositories (ROAR) lists a total of 3612 institutional repositories. However, in LIS domain most of the IDRs are cross-institutional i.e. these repositories allow submission of scholarly materials globally. Presently, 1685 open access repositories are OAI/PMH compatible in OpenDOAR among 2616 repositories (as on dated 19thMarch, 2014). Open DOAR repository includes 1554 multi-disciplinary subject fields in 2163 institutional repositories, 283 Disciplinary, 95 aggregating and 74 Governmental repositories. New services arise when the conditions are favorable. The emergence of the low-cost hardware, standardized software tools, open source software, open interoperability standards, low-cost communication devices, cheap storage devices, and distributed information system (open access journals, open access repositories, subject gateways) provided ideal ingredients for the development of a localized single window search interface for open access repositories (Chudnov, 1999). These single window search services can harvest metadata from different repositories that are open and compatible with interoperability standards. In such a system users can perform search centrally, display metadata of a resource from local server and retrieve full-text resource from the original server (may be anywhere in the world). As these services can be tuned to harvest metadata selectively, it may help in reducing cross-disciplinary semantic drift during search and retrieval.