DefiningImageAccess/Standard/METS
From ImageWeb
| DefiningImageAccess/Standard/METS | |
|---|---|
| Link:=http://www.loc.gov/standards/mets/}} | |
| Status:=Active}} | |
| JISCProject:=False}} | |
| Focus:=Describing Digital Objects}} | |
| ReferencedBy::}} | |
| [[PublishedBy::{{{PublishedBy}}}]]}} | |
| UsedBy::widely used by the libraries in the States}} | |
| DescribedBy::}} | |
| Partner::Network Development and MARC Standards Office of the Library of Congress}} | |
| [[Contact::{{{Contact}}}]]}} | |
METS
http://www.loc.gov/standards/mets/
A Library of Congress activity, Metadata Encoding and Transmission Standard: "The METS schema is a standard for encoding descriptive, administrative, and structural metadata regarding objects within a dia:tag:=digital library, expressed using the XML schema language of the World Wide Web Consortium. The standard is maintained in the Network Development and MARC Standards Office of the Library of Congress, and is being developed as an initiative of the Digital Library Federation." (from the web page)
There is a high-level introduction here: http://www.loc.gov/standards/mets/METSOverview.v2.html. Thie introductory paragraph makes a compelling case for the importance of metadata for managing digital works, compared with printed works.
This appears to be a fairly comprehensive standard for dealing with, as it says on the tin, "descriptive, administrative, and structural metadata for a digital library object, and for expressing the complex links between these various forms of metadata". Subject or content specific metadata is not covered, but there is a behavioural section, which "can be used to associate executable behaviors with content in the METS object". The examples show some image technical metadata elements that may be included in the Administrative Metadata section.
Thus, it seems that no information usable for reasoning over domain concepts is available, though it may be that the METS framework can incorporate domain-specific extensions.
METS is implemented in XML, and appears to use URIs (possibly in the form of URNs) so I imagine a mapping to RDF would be feasible, if not already defined. Metadata is grouped into sections within the document header; it seems that arbitrary metadata vocabularies or namespaces may be used (this leads me to wondering that metadata grouping is represented in the document header rather than as properties or classification of the metadata terms themselves). Complex hierarchical structures of data streams can be represented, as well as the existence of arbitrary hyperlinks between them. Data stream content may be denoted by reference or by direct inclusion in the METS object.
The importance of METS to our work is probably somewhat dependent on the amount of take-up it has or is likely to garner among academic repository operators and commercial publishers. I think that LoC backing gives it a fair amount of impetus in the US academic and Web communities.
METS, MPEG-21 and ORE
Coming back to review this at the end of the project, it seems to me that METS is a very competent and complete encapsulation framework, which is not obviously overburdened with mandatory detail, and I find myself wondering what it is that later projects such as Pathways, MPEG21-DID, ORE provide that is not already included in this established specification.
I also notice (from http://www.fedora.info/documents/master-spec-12.20.02.pdf) that Fedora uses METS as its internal storage format.
More comments are included in the page about ORE (DefiningImageAccess/Project/ORE).
Import domain-specific metadata into FlyTED EPrints using METS
This was investigated as part of the paper writing for the ESWC2008.
The METS header is able to contain arbitrary data streams~(including RDF) either by value or by reference; and OAI-PMH has native support for harvesting METS metadata. A local harvesting script could then use the OAI-PMH protocol to harvest the RDF format metadata held in the METS header. This is a more generic approach because it uses a standard metadata schema METS and enables third-parties to harvest semantic-rich metadata without acquiring any prior knowledge.
