3.2 – A registry of aid information
3.2.1 – Metadata for datasets
0 Such a registry would require some metadata describing each resource. Aid information is diverse and can range from geospatial information to budgetary data. Developing a registry would require specifying the lowest common denominator of metadata which will be applicable to any given item.
- Short name (as a unique identifier and for URL)
- Title/name of resource
- Download URL
- License/legal status
0 Other metadata could be added as required or optional fields – preferably adhering to some common metadata standard such as Dublin Core. For example, for datasets pertaining to a particular country, a country field could be added. For geospatial data metadata may include a bounding box of the region the data pertains to. For financial records pertaining to a certain time period, this information may be specified. The OECD has published a report and details of a metadata scheme for publishing datasets that could be informative here 1.
3.2.2 – Ontologies and tags: user driven resource description
0 Tags are a useful way of describing large collections of diverse knowledge resources. Many popular web services such as Flickr, Delicious and Twitter have found tagging to be useful to help their users explore and navigate the large amounts of content they produce.
0 In the past knowledge resources have been described according to some ‘ontology’ – a fixed, predetermined range of fields used to describe the resource, including names, dates, descriptions, relationships between items being described and so on. By contrast ’social’ web services often allow anyone to add any tag, or keyword, to describe a given resource. These keywords do not need to be in accordance with some pre-defined scheme.
0 For example, on the photo website Flickr, users can add appropriate tags to describe photographs, which can be anything from the names of people, places, events or objects depicted, observations about the subject matter, and so on. Whereas a photo catalogue with a fixed set of fields may only contain information in accordance with these fields (perhaps including the name, date and photographer of a given picture), social tagging arguably allows a much richer spectrum of description. These tags facilitate with the discovery of the items they describe – for example by allowing people to search for all the photographs tagged with ‘red sky’, taken in Reykjavik, or depicting a certain brand of car.
0 Many suggest that tags are more flexible and dynamic than a pre-specified ontology – as they can be added, updated or revised ‘on the fly’. We think there may be significant benefits to allowing users to tag, or otherwise describe, information resources related to international development.
3.2.3 – Linking, mirroring and storage
0 The registry would contain links to documents and datasets containing aid information. This material would be stored externally – either on the websites of the organisations publishing the information, or by third parties who specialise in storing such material (such as the Internet Archive or Talis’ Connected Commons). As the material would be open, others would be free to deposit it in external archives, mirror (or host copies of) it elsewhere, or to aggregate it with other material.
3.2.4 – Making the registry open
0 We recommend that the information in the registry should be published under an open license so that others can build on it and integrate it with other resources 2. To give a recent example of where this has been useful, in July 2009 the Sunlight Foundation announced a project building on the US Government’s data.gov website 3.