Interesting discussion that argues against four assumptions that underlie the absence of inquiry into how metadata should be created: the assumption that mediation by controlling authorities is undesirable, that rigorous metadata creation is too time consuming, that only authors and users of resources have the necessary knowledge to create meaningful metadata, and that metadata can be generated or resolved by machine. Their argument is based on what can only be called a series of failures of attempts to generate metadata: spelling errors, variant spellings for people's names, varying resource titles, classification errors and inconsistencies, and even variant date formats. This summary of the OAI Arc (federated search) project is typical (if wordy): "Even extensive interventions during the metadata conversion phase could not prevent the negative impact that poor metadata quality has on the search and linking facilities..."
Today: 5 Total: 5 [Direct link] [Share]
