• Institutional workflows: Imperial College
  
• College Professional Web Pages database:
  1. Personal contact details, image, publications, teaching & research commitments & interests.
  2. Generated by feeds from corporate databases (H/R, Library etc) and freestyle editing in College Web CMS (based on Oracle Portal)
  3. Connects to SPIRAL (learned journal repository using DSpace)
• For FOAF experiments, link PWP database into Oracle 11G database with 11G XML SDK and Semantic features (RDF repository & SPARQL)
• XSQL queries against relational database return results as well-formed XML.
• XSLT run in Oracle Internet Application Server (J2EE container + Apache HTTPD Server) transforms results from XSQL queries into FOAF & DC.
• J2EE Container and Oracle iAS have extensive cacheing to enhance performance.
• Further work
  1. Extend to rest of PWP data (books, events, meetings etc)
  2. Investigate feeds from publishers, e.g. Ovid/Medline who already provide peer-reviewed author keywords/keyphrases: express these in SKOS.
  3. Text-mine papers locally where keyphrases don't exist or against multiple vocabularies for interdisciplinary work e.g. bioengineering.
  4. Improve system architecture (move to pure XML and then RDF databases)
  5. Pick up RDF feeds from other systems

---

© H. S. Rzepa, 2008.