Searching in online documentation. Metadata are an option.

To find information quickly and easily is a vital quality feature in technical documentation. In online documentation, such as online help, the following ways of searching and finding are most commonly available:

• Table of contents
  Users search for information in a directory which depicts the headings as well as the hierarchical structure of all chapters. Access to information depends on the clarity of this structure and on the semantic unambiguity of the headings.
• Index
  Users search for information in a directory of keywords. However, to successfully find information, the users have to be able to match keywords to relevant issues.
• Full-text search
  Users search the entire content of the documentation for an arbitrary term. All topics that include this term are displayed in the search results. The quality of the search results tends to be unsatisfactory, as the list of hits can become very long.

Unfortunately, in large-scale documentations, none of these methods delivers first class results. Furthermore, the search is one-dimensional and does not represent any relations. Thus, the currently available search methods are both time-consuming and imprecise.

What are metadata?

It would be ideal, if information could be complemented by additional information describing content and easing access to the information users are looking for. This kind of additional information is called metadata. The use of metadata aims to identify, assess and manage knowledge in or about modules.

Example: Library catalogues

A simple case of the use of metadata is illustrated by traditional library catalogues. They contain the following metadata: author, year of publication, publisher, title, number of pages, and more. The special feature of metadata in a modern environment, such as the internet, is not only that you can search for a metadatum, but that several metadata can relate to one another. That way, it has become conceivable to search a library catalogue for audio books published by a specific author in the year 2012.

For large large-scale technical documentation, this development lags behind. The extent of available information is getting larger and larger, and sometimes, heterogeneous groups (customers, developers, writers) complement the information, for example in a wiki. To solve this predicament, companies have to depict information with predetermined terminology and have to develop metadata specifically for their technical documentation and other resources.

For technical documentation, a special classification of metadata has been developed. (Drewer, Petra / Ziegler, Wolfgang, (2011): Technische Dokumentation, Würzburg, Vogel Buchverlag, S. 364-366) The following example from the documentation of a coffee machine shows a number of relevant metadata:

• Life cycle of the documentation
  State: <In progress>
  Creator: <C5059190>
  
• Product-related, intrinsic
  Component: <Coffee filter>
  Product: <AS 500XY>
  
• Product-related, extrinsic
  Type: <Espresso>
  Manufacturer: <Coffee XY, München-Frankfurt>
• Information-related, intrinsic
  Type of information: <Instruction>
  Action: <Refill>
• Information-related, extrinsic
  Publication medium: <Print>
  Information product: <User manual>

The classification methods and the structure of metadata have to be elaborated carefully and implemented consistently.

Searching metadata in special portals

There are special index-based search machines, for example Lucene, that search the published documents for the metadata. The search is limited to certain areas and the quality of the search results is improved. In case of the coffee machine, users could search for instructions applicable for espresso machines and including a topic on coffee filters.

There are a number of big companies which already offer search functions based on metadata, using portals established specifically for such tasks.

Further information:
Content Delivery Portal
Intelligent Content
Metadata