Difference between revisions of "ac:metadataLanguage"
m (Created page with "{{Concept |label=Metadata Language |is defined by=ac:metadataLanguage |concept type=property |definition=Language of description and other metadata (but not necessarily of the...") |
m |
||
Line 4: | Line 4: | ||
|concept type=property | |concept type=property | ||
|definition=Language of description and other metadata (but not necessarily of the image itself) represented in ISO639-1 or -3. | |definition=Language of description and other metadata (but not necessarily of the image itself) represented in ISO639-1 or -3. | ||
− | |notes=This is NOT | + | |notes=This is NOT [[dcterms:language]], which is about the resource, not the metadata. Metadata Language is deliberately single-valued, imposing a requirement that multi-lingual metadata be represented as separate, complete, metadata records in which also the language-neutral items appear. Consumers can re-combine records by identity of multimedia Resource IDs (which are highly recommended to supply). In the face of metadata records of several languages for the same resource, this normative document offers no guidance as to how to disambiguate what language applies to what metadata. To disambiguate this, the metadata provider can provide a separate AC object for each different MetadataLanguage, each such AC object referring to the same multimedia resource Identifier. This comes at a cost of repeated data for the "language neutral" metadata items, but the alternative is to have a complex hierarchical structure for an AC object. Nothing in this document would prevent an implementer, e. g. of an XML-Schema representation, from providing a fully hierarchical schema and disambiguating that way. Users of a particular implementation should consult documentation specific for that implementation on this point, but it should always be correct, if convenient, to provide one metadata record per language. |
|hidden notes=Crosswalk: NBII = Language of Cataloging; KeyToNature = {{NS k2n | Metadata_Language }}; DublinCore = {{NS dcterms | language}} referring to a "metadata resource".; Morphbank = "English" | |hidden notes=Crosswalk: NBII = Language of Cataloging; KeyToNature = {{NS k2n | Metadata_Language }}; DublinCore = {{NS dcterms | language}} referring to a "metadata resource".; Morphbank = "English" | ||
}} | }} |
Revision as of 15:05, 4 October 2012
Metadata Language
|
Search for values |
Notes: This is NOT dcterms:language, which is about the resource, not the metadata. Metadata Language is deliberately single-valued, imposing a requirement that multi-lingual metadata be represented as separate, complete, metadata records in which also the language-neutral items appear. Consumers can re-combine records by identity of multimedia Resource IDs (which are highly recommended to supply). In the face of metadata records of several languages for the same resource, this normative document offers no guidance as to how to disambiguate what language applies to what metadata. To disambiguate this, the metadata provider can provide a separate AC object for each different MetadataLanguage, each such AC object referring to the same multimedia resource Identifier. This comes at a cost of repeated data for the "language neutral" metadata items, but the alternative is to have a complex hierarchical structure for an AC object. Nothing in this document would prevent an implementer, e. g. of an XML-Schema representation, from providing a fully hierarchical schema and disambiguating that way. Users of a particular implementation should consult documentation specific for that implementation on this point, but it should always be correct, if convenient, to provide one metadata record per language.