IMRC - Indexing Federal Government Web Pages: Guidelines for the Development of an Indexing Policy

Indexing Federal Government Web Pages

Guidelines for the Development of an Indexing Policy

Final version of: July 12, 2002

Table of Contents

Introduction
Reasons For an Indexing Policy
Standards
Policy Issues
Level of Specificity
Sources of Vocabulary
Official Languages
Training
Quality Assurance
Risk Assessment

Introduction

The present indexing policy recommendations apply to the Common Look and Feel metadata element: Subject, expressed as <dc.subject>.  This document serves as a guideline for the development of an indexing policy for individual departments. It is highly recommended to develop a local policy and to use local expertise such as:

• CLF Centre of Expertise members;
  
  
• GOL Metadata Working Group members;
  
  
• Departmental Librarians.

Reasons For an Indexing Policy

1. Under Common Look and Feel (CLF) Standard 6.3  (http://www.tbs-sct.gc.ca/clf-nsi/inter/inter-06-tb_e.asp#6.3), all Government of Canada (GoC) Web sites must adopt five core metadata elements for the description of Web resources including <dc.subject>.
   
   
2. The purpose of indexing is to provide an easier and more precise access to intellectual content.
   
   
3. Poor indexing practices degrade information retrieval and have serious cost implications.
   
   
4. A well-designed indexing policy will ensure consistency of access to all government information resources.

Standards

Indexing policies should be developed and used in conjunction with:

1. TBITS 39: Treasury Board Information Management Standard, Part 1: Government On-Line Metadata Standard
   http://www.tbs-sct.gc.ca/its-nit/standards/tbits39/crit391_e.asp
   
   
2. TBITS 39: Treasury Board Information Management Standard, Part 2: Controlled Vocabulary Standard
   http://www.tbs-sct.gc.ca/its-nit/standards/tbits39/crit392_e.asp

Policy Issues

1. The larger the number of documents to be indexed, the greater will be the requirement to establish priorities among document types to be indexed.
   
   
2. In general, index as many pages as possible, but index main pages, such as welcome and home pages first.  In addition, index pages containing:
   
   
   • Programs or services in high demand (especially online services);
     
     
   • Information about entitlements, rights or obligations;
     
     
   • Information about dangers or risks to health, safety or the environment;
     
     
   • Information required for understanding major new priorities, laws, policies, programs or services.

The TBS Metadata Standard (TBITS 39.1), Section 3 provides a complete list of resource types to which metadata (including index terms) should be applied.

http://www.tbs-sct.gc.ca/its-nit/standards/tbits39/crit391_e.asp

Level of Specificity

• To represent a concept or a subject, be as specific as the list of authorized descriptors allows you to be.
  
  
• Use as many authorized descriptors as needed to fully describe the contents of an information resource.
  
  
• More than one subject descriptor will be needed to describe most resources.

Sources of Vocabulary

As specified in TBITS 39.2, the default source of controlled subject vocabulary is the Government of Canada Core Subject Thesaurus (CST)

http://en.thesaurus.gc.ca/these/thes_e.html

This means that in the absence of a suitable, registered vocabulary, the CST must be used. If your organization has developed its own controlled vocabulary for use in indexing, then it must be registered with the National Library of Canada (NLC).  Another option is to select a vocabulary from the list of registered vocabularies on the NLC site.  For more information, consult the list on the NLC Web site.

http://www.collectionscanada.ca/8/4/r4-281-e.html

Official Languages

Documents are indexed in the language of the resource.

Training

• Indexers may be recruited from a number of areas and bring a wide range of skill levels to the required tasks.  Indexers may include:
  
  
  • Authors;
    
    
  • Content Developers;
    
    
  • Librarians;
    
    
  • Professional Indexers;
    
    
  • Webmasters.
    
    
• Various levels of training will be required depending upon the skill levels of particular individuals.
  
  
• Some consideration will have to be given to the language skills of indexers because indexing must be done in the language of the resource.

Quality Assurance

• A process of quality assurance should be developed and implemented to support the indexing process, in order to ensure that a particular concept or subject will always be represented in the same way.
  
  
• Information managers should be accountable for consistency and accuracy
  
  
• Departmental librarians are a good source of expertise in this area and should be consulted.

Risk Assessment

Failure to develop and implement a policy for indexing and the application of metadata may result in the inconsistent and inappropriate population of the relevant metadata elements.  This would seriously affect the efficiency and quality of information retrieval.  Indexing and metadata application are likely to have significant costs associated with them whether the work is done properly or not.  Revision of incorrectly or inadequately applied metadata will increase costs in the long term.