UKOLN International Repository Workshop: Citation Services

First Report:

• Currently small number of commercial service providers is dominant in this field. Are we evolving repository services [to accommodate the existing systems], or revolutionising them?
  
  
• Since citations drive national funding, systems need to be trusted auditable and open.
  
  
• Citations relate authors and ideas, and help connect concepts together; they provide literature ranking, and larger scale analytic services across literature.
  
  
• International coordination: existing infrastructure of loosely coupled repositories can be foundation of robust scalable solution.
  

Second Report:

• The group is producing no large plan and manifesto, but is going back to basics.
  
  
• "Handshake" meant different things to different people; there are limitations to the metaphor.
  
  
• There will be group activity, with two foci: business and technological.
  
  
• Recruitment of content needs to happen outside repository established space, including through desktop bibliographic tools such as Zotero.
  
  

Third Report:

• There is a huge variety of presentations of citations, and there are partial solutions specific to communities.
  
  
• Model how to deal with citations: Isolate references from papers, and then extract reference data, and interpret it, from varying citation schemes.
  
  
• For repository to be active in this without overconsuming resources, the repository shall be made responsible to hand on to external services the list of references extracted from their items (papers).
  
  
• Plan of action:
  
  
  
  • Establish test bed of references, out of what repositories find interesting.
    
  • Create repository API, repository plugin, OAI PMH profile.
    
  • JISC developer competition to develop toolkits.
    
  • Then liaise with e.g. Crossref and establish collaboration: the commercial bodies already have such services.
    
  • Then create a reference item processor as an external service, decomposing references into constituent data.
    
  • Then build services like Citeseer and Google Scholar—or use those existing services, if they will collaborate.
    
  • Then build exemplar GUI end user services, e.g. trackbacks, visualisations.
    
  • Liaising with publishers important but not a dependency for remaining tasks.