Metadata Infrastructures and Operations

1 Introduction

Metadata are only some lego blocks which can be used to build datasets to meet the functional requirements of different applications. How to use these blocks when we face an overwhelming blocks? Such issues include how to get the correct block, how to update it when a block changed and how to integrate blocks from different metadata standards. Fortunately, some pioneers have done some for the goal. Here, we describe what they have done on the metadata infrastructures and operations.

2 Metadata Infrastructures

Li Qi has presented a unified metadata information management framework for digital city. He presents a metadata management architecture used to facilitate resource discovery in the digital city's information resource network. His work addresses the metadata management architecture as well as three core operations, metadata standards mapping, metadata uploading and metadata querying.

Here, we summarize how to solve the three issues:

• Metadata standards mapping

  Generally, we have to use various metadata standards in the information resource, this bring a great difficulty for the metadata management, retrieval, and locating. Metadata standards mapping is very help.
  Currently, standard mapping is largely performed manually by domain experts. It is a time-consuming and tedious process. To assist it, a graphical interface is developed. The user can adjust the proposed mapping by simply removing or adding lines connecting the elements of two standard schemas.

• Metadata uploading

  Metadata uploading is like metadata integration. The metadata storage structure used in the paper has two layer, central node and sub-node.
  A typical process for metadata upload is as the following 2 steps:

  1. The sub-node harvests detailed metadata from the original data and stores it in the local content metadata repository.
  2. If the metadata standards of the sub-node’s and the central node’s are the same, the sub-node extracts the core metadata from the stored content metadata and pushes it to the central node.

  If they are not the same, the sub-node first extracts the core metadata from its content metadata. Then it converts the core metadata into the central node standard compliant metadata, using the metadata standards mapping stored at the local node. Next, it pushes the core metadata up to the central node.

• Metadata querying

  The central node is the unified portal for the metadata retrieval. The users input the query conditions to the metadata retrieval service at the central node. The server searches all the core metadata stored in the central node metadata repository. The user can set further query conditions to ask for the detail information with the results returned. The metadata service then forwards these retrieval conditions to the related sub-nodes to get the content metadata. The sub-node can also publish metadata and process metadata query as a complete server if needed.

3 Metadata Operations

In order to management heterogeneous metadata in different application systems, SHIHN-YUARN CHEN proposed a XML-based Metadata Management and Manipulation Framework (XMF for short).

XMF contains three main layers: storage layer, constraint layer and metadata manipulation and presentation layer. Storage layer stores metadata. The constraint layer handles the tasks of metadata retrial according to the given constraints, such as a keyword, or a unique id of an object. The metadata manipulation and presentation layer can transform the presentation from XML format into others, such as HTML. The transformation can be achieved by CSS, XSLT and programs.

For metadata retrieval, it uses key words. For the metadata update, it distinguishes tow types: metadata field value update and metadata filed schema update. Metadata field value update means that the data of a field is modified. Metadata field schema update means that the schema of metadata of one type of objects changes.