Metadata - Fit for Purpose
About This Document
This briefing document describes the issues to be considered when choosing and using metadata.
Why Use Metadata?
Metadata cannot solve all your resource management and discovery problems but it can play an important part in the solutions. Since time and effort is needed if metadata is to be used effectively, it is vital to look closely at the problems you wish to address.
Do you want to allow resources on your Web site to be found more easily by search engines such as Google? Or perhaps you want to improve local searching on your Web site? Do you need interoperability with other projects and services? Maybe you want to improve the maintenance of resources on your Web site.
While metadata has a role to play in all of these situations, different approaches will be needed to tackle each type of problem. And in some cases, metadata may not be the optimal solution; for example, Google makes limited use of metadata so an alternative strategy might be needed.
Identifying the Functionality to be Provided
Once you have clarified why you want to use metadata, you should identify the end-user functionality you wish to provide. This will enable you to define the metadata you need, how it should be represented, and how it should be created, managed and deployed.
Choosing The Metadata Standard
You will need to choose the metadata standard which is relevant for your purpose. In many cases this will be self-evident. For example, a project that is funded to develop resources in an OAI environment will need to use the OAI application, while for a database of collection descriptions you will need to use collection description metadata.
Off the Shelf or Custom Fit?
Some metadata can be used without further work - for example, MARC 21 format in library management system cataloguing modules or entries in the Cornucopia and MICHAEL collection description databases.
Other metadata requires decisions on your part. If you are using Dublin Core, you will need to decide whether to use qualifiers (and if so which) and which elements are mandatory and which are repeatable.
Managing Your Metadata
It is important that you think about this at an early stage. If not properly managed, metadata can become out-of-date; and since metadata is not normally displayed to end-users but processed by software, you won't be able to check visually. Poor quality data can be a major obstacle to interoperable services.
If, for example, you embed metadata directly into a file, you may find it difficult to maintain the metadata; e.g. if the creator changes their name or contact details. A better approach may be the use of a database (sometimes referred to as a metadata repository) which provides management capabilities.
Example Of Use Of This Approach
The Exploit Interactive e-journal was developed by UKOLN with EU funding. Metadata was required in order to provide enhanced searching for the end user. The specific functionality required was the ability to search by issue, article type, author and title and by funding body. In addition metadata was needed in order to assist the project manager producing reports, such as the numbers of different types of articles. This functionality helped to identify the qualified Dublin Core elements required.
The MS SiteServer software used to provide the service provided an indexing and searching capability for processing arbitrary metadata. It was therefore decided to provide Dublin Core metadata stored in <meta> tags in HTML pages. In order to allow the metadata to be more easily converted into other formats (e.g. XHTML) the metadata was held externally and converted to HTML by server-side scripts.
A case study which gives further information (and describes the limitations of the metadata management approach) is available.
Managing And Using Metadata In An E-Journal, QA Focus briefing document no. 1, UKOLN, <http://www.ukoln.ac.uk/qa-focus/documents/case-studies/case-study-01/>
