OAI - Open Archives Initiative
The open archive initiative deals with the dissemination of metadata and connection of electronic repositories. An institution publishes its catalogue via OAI-PMH so that others can search it. On the one hand there are providers who provide the metadata of what they have in their repository, on the other hand are the harvesters who provide a search interface to metadata repositories.
Providers
Although you can use several types of metadata, unqualified Dublin Core should be used first. Moreover, the metadata does not have to be stored in DC, but could be converted to it upon request by the harvester. But we consider using DC anyway.
eSciDoc and formerly eDoc are capable of providing OAI-metadata, how to enable OAI-PMH for eSciDoc is described on the escidoc-Wiki.
The Data Provider is a server that can process the 6 OAI-PMH requests. PMH means Protocal for Metadata Harvesting.
Harvesters
Harvesters are Service Providers. They can search an OAI-metadata repository for you. Harvesting is not crawling: the metadata are retained in their original form, which makes sense. To make oneself known to harvesters, one should register the repository.
Search capabilities
According to eSciDoc, you have advanced search capabilities via (in the case of eSciDoc) CQL (Common Query Language): Boolean, String, Phrase, Fuzzy and Near.
Register your repository
The existence of the repository can be made known by registering the repository on the OAI site. It is required that the repository conforms to some rules:
The requests that the repository gives out are to be in XML and they have to comply to an XML schema. And it is helpful if the metadata output is in oai_dc, that is, dublin core.
Every metadata record has to have a unique identifier. It is only mandatory that the identifier is unique in the repository itself. However, it is also possible to turn this into an OAI identifier in the OAI namespace.
