There have been two reports relating to metadata released recently that I’ve been meaning to read and blog about: OCLC’s What We’ve Learned from the RLG Partners Metadata Creation Workflows Survey and DLF’s Future Directions in Metadata Remediation for Metadata Aggregators. However, I’m not going to get a chance to do more than skim these for a while – so while they’re fresh here’s the links.
What We’ve Learned from the RLG Partners Metadata Creation Workflows Survey
There’s an interesting comparasion of tools and people used to create MARC and non-MARC materials.
Of particular interest is how little libraries seem to involved in ‘educational’ metadata – 8 out of 78 respondents are invovled in creating metadata for learning objects. Now I know there’s a world of difference between learning materials with educational metadata and learning object metadata but looking at the metadata standards being used, educational description still seems to be out of scope. It is also of note how little of the respondents metadata is currently being pushed to/ used by newer forms of exposure – web2.0 tools and SRU.
Future Directions in Metadata Remediation for Metadata Aggregators
Many repository and digital library services operate on the premise that exposing their metadata will enable information about thier content to be made available in larger resource discovvery services that aggregate metadata from many sources. Such services provide a valuable service but often have well documented problems with variation in the metadata which they harvest. Whether the metadata is of poor quality or simply designed to support local needs without consideration of a wider context, aggregated metadata needs to be cleaned and otherwise processed to provide a better discovery service. This report examines key services/ features that a aggregated search service would hope to provide and for each documents how metadata supports that service, what tools exist to ‘fix’ harvested metadata, what tools are desired and provides a bibliography and comments. As such this report should represent an overview state of the art (and I really need to read it soon…).