DefiningImageAccess/RecommendationNotes
From ImageWeb
Contents |
Recommendations for Image Repositories
WORK IN PROGRESS: this is all very drafty and very incomplete
Persistence of URIs
Once information is published at a given URI, try to keep it available at that URI. If URIs change, trails through the web are broken, and important image collections may become orphaned from related collections. For machine accessed data, the concerns of link breakage are probably more serious than for human-readable pages, as software agents probably won't have access to alternative mechanisms to find the information that are available to human browsers (e.g. Google for a title).
See: http://www.w3.org/Provider/Style/URI
The JISC URI http://www.jisc.ac.uk/index.cfm?name=wg_images_home is an example of this kind of breakage - the indicated information is no longer available using that URI.
Metadata
Role of domain specific metadata
Most institutional repository services focus on generic metadata - general, structural and presentation metadata that is not specific to the content of the digital objects. Structural metadata comes nearest (e.g. image format, size, etc.). For usefully linking research image data across repositories, we need the metadata to tell us what an image is about; e.g. for gene expression images, details like organism, strain, expressed gene(s), anatomical region, etc., without which it is not possible for another researcher to say what the image might indicate.
Thus, it seems that domain-specific metadata may have a role that needs to be handled differently from generic metadata. There are also repository management issues, in that domain metadata may be continually changing and adapting as new ideas are developed. It could be that, in repository terms, such metadata must be treated more like source data than metadata.
See:
