News | about | Search Language: DE | EN | FR
 Related Article: Growing Flora of Metadata Standards

European Treasury Browser
ETB Handbook
•  ETB Controlled Lists of Vocabularies
•  ETB Datamodel: Minimum set of Quality Descriptors
•  ETB Tech Update: Network
•  Inside ETB Tools: Exchanging Metadata
•  Introduction to Collection Level Descriptions
•  New ETB tools
•  Related Article: First Step Strawman
•  Related Article: Growing Flora of Metadata Standards
•  Related Article: Hype about metadata registries
•  Related Article: Learning Technologies Workshop - CEN/ISSS
•  Related Article: Metadata Observatory
•  Related Article: Prof Erik Duval On MoU between LOM and DCMI
•  Repository owner, are you interested in joining the ETB-network?
Multilinguality and Thesaurus
Outputs and Documents
Search this area

Advanced Search

Print this page!
Tell a friend!
To get a reminder when this page is updated, please enter your email address here:

Other languages: [fr ]

This note is not so much about metadata per se, as about the kind of problems this vigorously growing flora of metadata standards poses for those who implement it.

Now, in general, there will be no problem implementors who have a homogeneous target audience and a very narrow scope of their service, but life is not as easy as that. Let us take a real-world example...

Metadata is used for describing any kind of data in order to make those data easier to find. Some metadata are absolutely essential, and a web browser just cannot do without them, for instance, the browser need to know the physical format of a retrieved resource or it cannot be rendered by the browser or select the appropriate plug-in or helper application. If such metadata are incorrect, it will generate an error in the browser, and in a sense, metadata is the data needed to make information out of what is otherwise just an disorganised digital blob of zeros and ones.

Apart from the very simplest types of metadata (like mime type and content size), one of the main problems of metadata is that there is a large number of standards available. Each of those standards have their own application area; that is, each of them are good for describing certain kinds objects. So, for instance, MAchine-Readable Cataloging (MARC) [1] has been used for decades in the description of all kinds of media in the library communities. It is also used by archivists for describing archival collection, but this community is now more and more turning to Encoded Archival Description (EAD) [2]. I could go on and on and give acronyms for metadata element sets, both established standards and new projects. In the education sector, we see the IMS Project [3], DC-Education and the UK Metadata for Education Group (MEG) [5]. There are more of these, both initiatives geared towards education or other sectors of society or industry.

Now, in general, there will be no problem implementors who have a homogeneous target audience and a very narrow scope of their service, but life is not as easy as that. Let us take a real-world example. Go to the US Department of Education and search for the acronym GILS [6]. All federal governmental agencies in the US are by law required to provide metadata for reports in the Governmental Information Locator System (GILS) [7]. You will find a few such records on their site, and also tens of thousands of them in a search service hosted by ERIC. It is an obvious advantage that governmental information of all kind is made available through a single metadata standard. But the same information, one could argue, should be available on public libraries, where they have to be search-able through MARC in a library OPAC and perhaps through IMS in educational databases. You already see the problem, I suppose: The problem of interoperability. The same data may need metadata for different purposes, for different situations or target audiences.

One of the first initiatives to address this problem was the Dublin Core Metadata Initiative (DCMI) [8]. The Dublin Core metadata element set is meant to represent the least common denominator of the more specialized, and also more complex, element sets used in specific communities. In addition, the people designing the DC had in mind the functional requirements for simple resource discovery on the Internet. The effort has been successful. Just about any detailed description should be able to 'down-grade' into DC, regardless of the source element set used to create it in the first place.

But down-grading has its price. Information is lost, and what has been down-graded can nerver be up-graded. And it has to be done with care, or the result might be erroneous or misleading. New ideas are now emerging where a more atomic approach is taken in the task of creating a metadata element set.

New views on metadata are now possible because of the development of new technologies such as XML [9] and RDF [10]. One such attempt is the Harmony project [11] which attempts at building a general framework for metadata vocabularies [12]. In the future, metadata elements will not be thought of as living in specific element sets. Rather, implementors will be able to build their applications using elements from different sources, called namespaces [13]. These new kinds elements, which are reusable entities made available in different namespaces designed for different needs, will if chosen with care, fit together just like LEGO pieces. The Harmony project is now building a metadata element thesaurus for testing this idea. Such a thesaurus will contain information about the relations between metadata word, such that Director, Author and Illustrator all have the same broader term, Creator. Subject has narrower terms Classification and Keywords. So in the future we might see one single service, which using a modular approach to resource description achieves:

  • Interoperability with library OPACS through MARC and library classification systems.
  • Having simple access interoperable with Internet search services through Dublin Core.
  • Use geospatial metadata for geographical retrieval of geographical teaching resources.
  • Allow searching based en educational metadata using IMS [8]

We are not there yet. And, when I think of it, noone has volunteered to create the metadata.


Author: Sigfrid Lundberg, Lub NetLab
Web Editor: Riina Vuorikari
Published: Tuesday, 30 Jan 2001
Last changed: Tuesday, 9 Oct 2001
Keywords: metadata, interoperability