This diagram offers a talking point around what we mean by content integration in a large organisation such as the University of Bristol and has been used recently in discussions with CIP stakeholders at the University.
Data in our system can be acquired from central systems within the University (such as Oracle databases holding staff and research data), external “services” such as funder websites, or via social software means – from University researchers, say. As there are multiple central systems at the University of Bristol, we note that a systems integration layer may well need to be present for the ‘local’ aspect of data acquisition.
The content integration layer is where we position our CIP pilot software in the main. It is responsible for interpreting the semantics that describe “what this is” for the content being integrated during data acquisition. We are using semantic web technology to perform the required functionality at this stage. We use it to map relational database schemas in use in the University’s central systems and to describe and therefore easily aggregate external content. Hence the CIP system can semantically interpret data according to such concepts as “researcher”, “funder”, “research output”, across all its heterogeneous data sources.
The presentation layer represents components that can provide flexible views onto all or part of the data offered via the content integration layer. These views could be customised for several purposes such as a departmental content management system, the external – public -, web-based view of University research, or via a toolkit for internal business intelligence activity and research management.
We understand that access control and data privacy are likely to be important concerns in a CIP-like system. Our project links with the Web Futures team’s Caboto project on this aspect. We are also developing OpenId support for CIP.