Organizing USGS information with consistent vocabulary

Peter N. Schweitzer (U.S. Geological Survey, Reston VA 20192)

This presentation is included as part of the short course "Designing information for the worldwide web" presented at USGS Publications 2003

Audience:

People engaged in the USGS publication process who are learning how to design information for the web. They will have learned the conceptual method of Information Mapping, Inc., including principles of relevance, chunking, and labeling.

Goal:

This presentation is intended to enhance peoples' understanding of labeling by relating it to the design and use of controlled vocabularies in catalogs and indexes within USGS.

Contents

1. Understand indexing with controlled vocabularies
   1. What are controlled vocabularies?

      Consistent collection of terms chosen for specific purposes with explicitly stated, logical constraints on their intended meanings and relationships. Includes classifications, formal thesauri, topic maps, subject headings, authority files, gazetteers, and ontologies.

   2. Why indexing isn't the same thing as labeling

      Labels shown on web pages need to be quickly recognizable with minimal additional study by the user; index terms specify how information chunks are related to other chunks and to other index terms. Labels might need to identify combinations of concepts that are conceptually distinct.

      Artificial example:

      Label: Fish in Green Lake
      Index terms: fish, lakes, lake fish, Green Lake
      

      Example from USGS web:

      http://toxics.usgs.gov/topics/attenuation.html
      Labels:
        Toxics Program Projects on Natural Attenuation
          science programs
            and
          contamination and pollution
            and
          biodegradation
            and
          contaminant transport
            and
          geochemical processes
        Headlines
          news services
        Natural Attenuation Remediation Related Activities
          contamination and pollution
            and
          remediation
        Fact Sheets
          reports
        Bibliographies
          bibliographies
      

   3. How well-constructed vocabularies help label relevant chunks well

      Labels can be drawn from controlled vocabularies to increase consistency and familiarity. By recognizing in your web site the concepts represented in the vocabulary, you may be less likely to present those concepts in a confusing way.

2. How controlled vocabularies fit into the USGS enterprise web
   1. Design of the USGS thesaurus
      1. Concepts, not words, constitute terms
         1. Arguments about specific terms go on forever
         2. Instead, focus on concepts
      2. Term relationships (BT, NT, RT, UF, SN)

         BT	Broader Term
         NT	Narrower Term Terms in hierarchy must have an is a relationship NT is a type of BT NT is a part of BT NT is an instance of BT
         RT	Related Term We do not describe the nature of the relationship beyond the notion "you might be interested in this" Ontologies and topic maps specify the nature of these relationships in detail and focus on them
         UF	Use For (relates non-preferred to preferred terms). Not synonymy, but put in the same bucket for search purposes Non-preferred terms are also called lead-in terms.
         SN	Scope Note explains our use of the term, may include definition

         Examples

         geologic contacts
           SN:   Plane or irregular surface between two types
                 or ages of rock; examples are faults, intrusive
                 borders, bedding planes separating distinct
                 strata, and unconformities. [Glossary of Geology, 4th ed.]
           BT:   stratigraphic sections
           NT:   unconformities
           RT:   stratigraphy
           UF:   contacts (geologic)
         
         geologic history
           SN:   Record (and inferred reconstruction) of the origin
                 and development of the Earth since its formation.
           BT:   Earth characteristics
           NT:   biostratigraphy
                 Earth history
                 lithostratigraphy
           RT:   geologic time scales
                 geology
                 paleontology
                 paleoseismology
                 stratigraphy
           UF:   chronostratigraphy
                 geohistory
         
         ecological processes
           SN:   Dynamic biogeochemical interactions that occur
                 among and between biotic and abiotic components
                 of the biosphere.
           BT:   biological and physical processes
           NT:   algal blooms
                 bioaccumulation
                 biogeochemical cycling
                 biological productivity
                 contaminant transport
                 dispersal (organisms)
                 ecological competition
                 ecosystem functions
                 eutrophication
                 extinction and extirpation
                 habitat alteration
                 migration (organisms)
                 pollination
                 succession (biological)
           RT:   ecology
                 population and community ecology
           UF:   environmental processes
           UF+:  ecological models
         

      3. Scope: broad and shallow
         1. Covers topics to which USGS contributes useful information directly
         2. Includes a few topics of collateral interest to our users if information on those topics is supplied by our partners
   2. Catalog as an application of the thesaurus
      1. Scope and design
         1. Important resources not part of publications warehouse
            Web pages dealing with
            • subjects
            • geographic areas
            • programs and projects
            • organizational units
            • software
            • services (IMS, database search)
         2. Head page of a document indexed, not lower pages
         3. Bibliographic information, index terms, description, contact
         4. Implemented in relational database
      2. Purposes
         1. Test the thesaurus as an indexing tool
            • Find concepts not represented in thesaurus
            • Clarify ambiguous terms
            • Identify lead-in terms
         2. Provide a way to find information by topic, science, method, place
            • Pick terms from list rather than type in
            • Show scope and context for each term
            • Show documents matching given term or more specific terms
      3. Relationship to pubs warehouse
         1. Catalog demonstrates hierarchical browse strategy
         2. Includes some non-pubs resources
   3. Metadata within HTML documents
      1. USGS web guidelines specify elements

         http://internal.usgs.gov/guidelines/usgs.webguide.html#meta

      2. Search engines may or may not see
         • Most commonly seen: description, keywords
         • USGS search engine can see more (customized by USGS)
         • External search systems don't know our thesaurus
      3. HTML syntax cannot show term relationships

         <meta name="keywords" content="ecological processes">

         No place to put BT, NT, RT, UF
      4. Consistency with catalog?
         1. Catalog can generate meta tags
         2. Harvest meta tags--possible
         3. Different purposes for terms: categorize vs characterize
            Example: Natural attenuation
   4. Use of other vocabularies
      1. Facets not included in thesaurus:
         1. Place names
         2. Geographic feature types
         3. Geologic time
         4. Stratigraphic names
      2. More details for included facets:
         1. Rock types and mineral names
         2. Biological taxonomy
      3. Authority lists:
         1. Publication series names
         2. Organizations
         3. People