Below is the uncorrected machine-read text of this chapter, intended to provide our own search engines and external engines with highly rich, chapter-representative searchable text of each book. Because it is UNCORRECTED material, please consider the following text as a useful but insufficient proxy for the authoritative book pages.
14 The goal of this research was to describe and assess the current state and future potential of the TRT. However, the TRT has a long history, and it is important to understand when and why it was created and how it has evolved to its current form. This history is best understood in the context of the development of thesauri, TMS tools, and the use of thesauri since the 1950s. The research team identified four generations in the history and evolution of thesauri, thesaurus architectures, and technologies. History of Thesauri and the TRT In Generation 1, the thesaurus was in book form for manual perusing. In Generation 2, it became a component of information management systems. In Generation 3, stand-alone TMS applications were introduced. During these first three generations, a thesaurus was understood to be a tool for information professionals in subject analysis, indexing, and search. These early generations aligned with the early and progressive developments of technology and the increas- ing but limited access of the public to those technologies. The fourth generation saw the impact of semantic technologies and artificial intelligence. In this generation, thesauri are seen as critical components of semantic architectures and as tools that stand on their own and are consumed by a larger community to navigate the new semantic world. Where does the TRT fit into this historical context? The TRT was originally developed from 1993 to 1996, prior to the introduction of standards such as ISO 25964 (see Figure 2-1). The TRT was born in Generation 2, as the need for vocabulary management grew in the field of transportation. As originally developed, the TRT was an elegant design, well suited to its purpose. The TRT was designed by one of the foremost experts in the field of thesaurus management and vocabu- lary control, Dr. David Batty. The TRT was designed to organize information into broad classes that could be searched-browsed in the TRID repositoryâsimilar to subject headings in library catalogs. Because TRID was developed in proprietary software, so too was the TRT component. In 2008, the TRT indexer interface and thesaurus management interface were embedded into the TRID application. In this context, they resemble the catalogerâs interface in an integrated library system. This approach puts the TRT firmly in Generation 2 as a component of another information system. Thesaurus Standards and Conceptual Model How does the TRT align with the current semantic world? The conceptual reference model derived from the ISO 25964 standard for thesauri (see Figure 2-1) suggests that there are three essential components of a standards-compliant thesaurus: ThesaurusConcept, ConceptGroup, C h a p t e r 2 The Transportation Research ThesaurusâHistorical and Current Context
the transportation research thesaurusâhistorical and Current Context 15 and Relationships. These components translate to thesaurus term, facet, or category, and fully specified thesaurus relationships. In the thesaurus standard, the central and essential element is the concept or term. Each term has a fully specified record that is used to manage that term. Designing around individual terms enables flexible definition of reports, search capabilities, multiple browse structures, the genera- tion of Key Word in Context (KWIC) and Key Word Out of Context (KWOC) structures, and management of the thesaurus at multiple levels. This architecture also enables the assignment of terms to multiple âgroupsâ or categories, which provides significant functionality, particularly for transdisciplinary domains such as transportation. Each term is linked to other terms through the specification of relationships. Relationships are structures that are attached to thesaurus terms. This model provides significant flexibility in representing the language of a domain. Areas of Variance from Thesaurus Standards How does the TRT vary from the ISO standard? The TRT varies from the standard concep- tual model in that the hierarchical relationships represented in the enumeration scheme are neither a ConceptGroup nor are they fully specified hierarchical relationships linking terms. A Figure 2-1. Model derived from the ISO 25964 conceptual reference model.
16 the transportation research thesaurus: Capabilities and enhancements ConceptGroup may be created by extracting all of the terms that are assigned to a TRT alpha- betically enumerated facet. However, this is not an existing feature of the TRT. Instead, the TRT does not have fully specified hierarchical relationships assigned to individual terms. Hierarchical relationships are interpreted on the fly for the TRT display from a translation of the alphabetical enumeration scheme. The TRT as a Component of TRID Search Why does the TRT vary from the ISO standard? The TRT varies from the standard because it was designed as a component of a search engine. Figure 2-2 describes the TRT as a schema that is part of the larger TRID search system. The conceptual model of a search system includes the following basic elements: bibliographic records, metadata repository, controlled reference structures, full-text index structure, hybrid index structure, attribute-based term index, query- processing algorithms, query-matching algorithms, search results display, and a search interface. Controlled reference structures help the searcher to select the ârightâ form of the term to search if they are searching by specific fields. In the TRID example, the TRT is a tightly integrated con- trolled reference source that supports attribute-based term searching. The TRTâs primary func- tion as a support to index-based searching explains why the TRT schema refers to index records and components rather than to thesaurus standard components. This context means that the focus is not on term records and their full functionality but on the specific use of those terms in metadata records to support TRID. Figure 2-2. The TRT database schema.
the transportation research thesaurusâhistorical and Current Context 17 Within the context of TRID, the TRT terms are linked through a well-developed and designed search index architecture. This is an important linkage point for thesaurus term records, includ- ing the linking of preferred (enumerated) and non-preferred (non-enumerated) terms. While this structure makes very good sense from a search perspective, it presents challenges for thesau- rus managers and users. From a standards perspective, this means that the focus is not on term records and their full functionality but on the specific use of those terms in the search index. Although the TRTâs database architecture may be efficiently designed from a search perspective (see Figure 2-2), this design does not support management of the TRT or its use in other applica- tions. The design also presents a development environment that would be costly to reconfigure in order to extend the thesaurus management capabilities that are currently available in commercial thesaurus management applications. The architecture and the functionality of the TRT are not consistent with Generation 3 com- mercial TMS applications. In addition, the TRT architecture imposes constraints on content that prevent the TRT from complying with current national and international thesaurus standards. These constraints present challenges for the TRT moving into Generations 3 and 4. Given the evo- lution of the external information and transportation environments, the TRT is now suboptimal in terms of scope and coverage, technology, governance, and architecture design. Despite their differences, the ISO model and the current TRT schema are not mutually exclu- sive. The goal of the TRT as a controlled reference source in TRID can be accomplished by creat- ing an ISO-compliant TRT. The elements needed to enable TRT-guided search in TRID would be effectively supported by an ISO-compliant thesaurus, and an ISO-compliant thesaurus would enable the TRT community to move content and use into Generations 3 and 4.
