Using Software to Teach Thesaurus Development and Indexing in Graduate Programs of LIS and IAKM

Bulletin of the American Society for Information Science & Technology, Aug/Sep 2005 by Zeng, Marcia Lei

The School of Library and Information Science (SLIS) at Kent State University offers the only ALA-accredited MLIS (Master of Library and Information Science) program in the state of Ohio. It is also the home of a new interdisciplinary program - information architecture and knowledge management (IAKM). To obtain the master of science degree, IAKM students must take the core program and elect one of the three concentrations: information architecture, information use or knowledge management. The IAKM program prepares information professionals to develop and manage information interfaces, products, systems and services for specific information ecologies.

I teach two courses related to thesaurus development and indexing. Indexing and Abstracting (LIS60649) covers manual and machine-aided indexing, creation of various types of traditional indexes, thesaurus construction and creation of sitemaps and site indexes. Knowledge Organization Systems: Taxonomy, Thesaurus, Ontology (IAKM60002) is a core course for IAKM students and a special topic course for LIS students.

Figure 1 shows the types of knowledge organization systems (KOS), arranged according to the degree of controls introduced (from natural language to controlled language) and the strength of their semantic structure (from weakly structured to strongly structured). It represents my visualized summarization of the Taxonomy of Knowledge Organization Sources/Systems (http://nkos.slis. kent. edu/KOS_ taxonomy.htm) adopted by the NKOS (Networked Knowledge Organization Systems/ Services) group based on Gail Hodge's article on KOS (www.clir.org/pubs/abstract/pub91abst.html). The marks in the figure indicate that assignments have been included in my IAKM class for the students to use and/or construct these types of KOS.

Students must understand not only how to construct KOS but also how to use KOS at various tasks, ranging from indexing, browsing and retrieval to knowledge management and information architecture design. Cases dealing with KOS applications are presented to let students connect what they are learning to the real world. These cases include government agency indexing systems, intranets, search engines, digital collections, digital learning environment, international trading and collaborative projects.

Several software programs have been introduced to the classes. They include the following:

* Thesaurus Management System, an in-house software package for thesaurus constructing and cross-thesaurus searching (http://circe.slis.kent.edu/mzeng/thesaurihome.html)

* Cindex, a commercial index-creation software package (http://161.58.146.10/home.ph)

* Protégé, an open source ontology editor and knowledge-acquisition system (http://protege.stanford.edu/)

* MAIstro, a commercial software suite for thesaurus construction and machine-aided indexing (www. accessinn.com)

* NASA thesaurus-based Machine-Aided Indexing (MAI), a tool available on the Web for machine-aided indexing process (http://mai.larc.nasa.gov/)

In the next section, I will compare the pros and cons of using some of them.

Pros and Cons of KOS Construction Software Used in Classes

Thesaurus Management System. Several years ago, I designed and created a software program with the assistance of a programmer. It has been available on the Web. I have used this software to check students' manually created thesauri. Students have used it to create their thesauri and to merge thesauri. The software allows online editing, including defining and changing terms and term relationships (facet/category, second language term, scope note and relationships such as USE, UF, BT, NT, RT, TT and CC). The strengths and weaknesses of the Thesaurus Management System follow:

Strengths:

* It is an open source.

* It is Web-based, allowing anyone to access it from any where at anytime.

* In addition to online editing, it is also designed for cross-thesaurus searching; therefore, it could be used for thesaurus merging and mapping.

* The established or imported candidate terms are automatically stored and provided in the editing process through the editing template's pull-down lists.

Weaknesses:

* It still has bugs when inputting long scope notes.

* It cannot process some special characters.

* There is no protection on the products. A thesaurus could be changed or deleted by anyone who enters the sytem's website.

* No statistics (such as word frequency) are available.

* The output is limited to the traditional thesaurus format, no XML (extensible Markup Language), RDF (Resource Development Framework, an XML schema) or OWL (Web Ontology Language) output.

* The experience has not been linked with any indexing process.

Protégé. Protégé is an open-source ontology editor and knowledge-base framework. The tool allows users to construct domain ontologies, customize data entry forms and enter data. It is also a platform that can easily be extended to include graphical components, media and various storage formats such as OWL, RDF, XML and HTML. Products like Protégé can be used to host thesauri as well, as illustrated by Wielinga's work, which employed Art and Architecture Thesaurus and Visual Resource Association (VRA) Core Metadata Categories (www.cs.vu.nl/~guus/papers/Wielinga01a.pdf). In my class, ontology creation is taught after students have learned different types of KOS, semantic relationships, encoding systems and metadata schemas. Teaching basic Protégé in my class has been a pleasant experience. The strengths and weaknesses of Protégé follow: