Centralisation or Federation?
Metadata Encoding and Transmission Standard (METS) Official Web Site) and the newly developed OAI-ORE (Open Archives Initiative Protocol – Object Exchange and Reuse) protocol for repositories. The federated model is based on sharing data and not duplicating entities, while the centralised model is based on duplicating and possessing data. The world is moving towards federated models where search engines visit repositories and extract what they wish. There is little point in copying data – the modern approach is to take the software to the data. It relies on it being public and Open – which Pubchem, WWMM and similar sites are. In WWMM we are developing a number of repositories where the data are new and unique to the repository and where the links will be passed to Pubchem for indexing and searching (there will also be local search facilities). OAI-PMH (Open Archives Initiative – Protocol for Metadata Harvesting – v.2.0) and OAI-ORE provide interfaces which allow search engines to extract metadata and data from repositories without human intervention. I am an advisor to the ORE project and will be meeting with the architects tomorrow – more later.=== PMR === The WWMM is a federated model while Chemspider is a centralised one. The WWMM is based on developing an agreed set of metadata for discovery and re-use based on Dublin Core, and more recently METS (