�

Why we need Metadata
Library is the traditional information repository to find most information needed. Library cataloguing is the method of organizing bibliographic information of library collections to facilitate their identification, location, access, and use. But Internet suddenly becomes a major mechanism for users connecting to a tremendous amount of digital resources directly from their own desktops. The Web itself becomes a huge free information database, a virtual library and also a path to acquire resources from other databases within the network.

Coming with this excitement are the phenomena that most of the Internet information is poorly organized, not stable. It is difficult for normal user to search or browse this huge wealth efficiently. There is an increasing demand to have a cataloguing mechanism for the Internet information. Different types of questions are raised when facing this challenge: are the methods of organizing the digital collections similar to those of organizing traditional library materials? Do we need new describing tools and techniques to catalogue digital resources? What type of standard we deploy? They will be created and maintained by professional librarians or normal users? What skills will be needed for this word and how could users get these skills?

Metadata could be the main if not the only method to answer those questions in the digital world. It helps provide needed descriptions for electronic resources, decrease users' search and discovery time, and what is the most important - increase the recall and precision of retrieval.

What is Metadata?
The most common definition of Metadata is "data about data." ALA Committee on Cataloging: Description and Access lists 27 definitions of Metadata from different groups and projects1. It's very common that most organizations start to create their own digital collections according to the specific types of structure or format of the resources inside. But currently there are strong demands on sharing those digital resources to save the time and investments. For users who want to retrieval and use those resources from different locations and services through a single interface, it's nearly impossible for them to know the specific retrieval requirements of each type of resource in advance. So there should be a mechanism that removes the barriers between different systems, provides common search terms, access points, and data structures, and also makes those differences transparent to end users.

Metadata is typically designed for describing the bibliographic information and data structure of resources among different systems to distinguish one piece from another, let user easily locate, access and transfer the data between. According to Prairie Village, the term "metadata" has been used since the early 1980s by the computer software and systems development community to describe the information required to document the characteristics of information contained within databases. But right now the resource described by Metadata is not limited to the document, it includes text, image, audio, video and any other format.

For a typical Metadata standard, it should have a complete elements set, each of which represents one property of the resource (normally called entity) described, with a structure that how these elements are organized, and a controlled vocabulary to define the name and value of elements. This sounds very familiar to librarians and cataloguers, since the MARC standard that is wide used in library community can be treated as a sample of Metadata. The concept of Metadata is being spread from library science to other domains such as computer science, and commonly accepted by all. Since the distribution of the first draft version of Text Encoding Initiative (TEI) Guideline in 1990, there have been a number of Metadata standards or schemas created covering a wide range of communities. Some of these metadata schemas are general, such as MARC or the Dublin Core, designed for electronic resources used in various disciplines. Other metadata schemas are dealing with specific discipline or domain, such as Government Information Location Services (GILS) and the . FGDC Content Standards for Digital Geospatial Metadata

The most common feature among these schemas is that they all use a set of defined elements to describe the properties of individual entity. As Vellucci indicated, there are three basic characteristics common to all metadata schemes:
�� 1. syntax
�� 2. semantics
�� 3. structure
To clearly identify each entity within the collection, each element can have one or more qualifiers to provide additional semantics to the values of elements, enhance the authority control to distinguish different entities.

On the other hand, the number of elements of metadata, how they describe the entity, and their structure are quite different. Besides the traditional description function, some schemas also include administrative metadata elements, structural metadata elements to manage the entity, the link between the entities, the access points and access rights to the entities. As an independent part from the entity described, the metadata elements can be stored together with the entity in the same file, or separately as a document in another physical location.