Structure of the Indexing Language
Suitability for the LEMoN Armoury * Type of Indexing Language and
Pre-coordinate Headings vs. Post Co-ordinate Retrievial
Forms of Terms
Precision / Recall * Specificity * Exhaustivity * Weight of Aboutness vs. Meaning
Structure of the Indexing Language
*Suitability for the LEMoN Armoury
The LEMoN Armoury Thesaurus (LAT) is intended to support an electronic database of product descriptions, which are used in our online catalogue as well as our internal training. The indexing language has been designed to meet the needs of several types of LEMoN Armoury employees: sword masters and apprentices, the Chief Information Manager (CIM), indexer, IT maintenance personnel, and sales and marketing staff (see Current Thesaurus Users for descriptions of each). Matching consistent terminology to products and processes will foster organizational knowledge and ensure high-quality customer service in taking and filling custom orders.
Although the thesaurus serves a relatively diverse group of users, it is not available to the general public and can therefore employ greater specificity and technicality with its terminology. Users of the thesaurus will be provided background training on weaponry appropriate to their position. Since the indexing language covers concrete objects that have been made and studied for centuries, some of the terminology used is unlikely to change. Nevertheless, both language and business needs change over time, and some ongoing maintenance of the thesaurus will be required. Especially important to the appointed indexer will be the careful and systematic addition of terms to cover other types of weapons produced by LEMoN Armoury. Multiple hierarchies and complex relationships will need to be explored as the thesaurus continues to grow.
NOTE: At all stages, the LAT has attempted to conform to the Guidelines for the Construction, Format, and Management of Monolingual Controlled Vocabularies (ANSI/NISO Z39.19-2005). Section number citations from this source are given throughout this structural analysis.
*Type of indexing language
Controlled vocabulary vs. natural language
As noted above, the intended users of the LAT will have some expertise in the field, appropriate to their respective positions and training in the organization. As a result, users should have a general sense of which terms will appear before beginning a database search. (Users may also develop familiarity with the thesaurus structure if it is used as a training tool.) Given these expectations, items in the collection will be indexed using only controlled vocabulary terms. The appearance of natural language descriptors will be restricted to entry (i.e., non-preferred) terms.
Dealing with Scottish swords and associated terms, the pilot version of the LAT covers only a handful of aspects most important to our users. These are: sword types, sword parts, swordmaking, sword wielding types, and sword performance characteristics. These aspects have been selected as the foundation for the prototype because the organizational focus is on the manufacturing of weapons – in this case, swords. The process and merchandising emphasis of our users mean that sword parts, design elements, and performance characteristics are more immediately important than are other aspects such as sword function. However, future expansion could include many other aspects relevant to sales and customer requests.
Additional considerations for language construction
Users of the LAT are expected to have some basic understanding of the existing collection. Nevertheless, user type varies within the organization, and not all those entering terms into the thesaurus will be handling the collection in the same way. In order to reduce possible confusion and enhance consistency, scope notes have been included where terms are vague, unusual, or imply technical knowledge.
The current prototype is mono-hierarchical, as it deals with only one branch of sword type (Scottish swords) in its present form. The LEMoN Armoury produces many types of weapons, however, and the thesaurus will eventually be multi-hierarchical when expanded. For instance, some characteristics now included as “sword parts” (such as “grip”) will apply to multiple types of “edged weapons” (such as axes and daggers). These relationships will need to be displayed in a form not exhibited by the current prototype.
In an attempt to glean solid technical terms from reliable sources, literary warrant [18.104.22.168] has been used to compose the major substance of the LAT. Literary warrant fomented the familiarization process and allowed term validation between multiple resources, including weapons encyclopaedias, history books, museum descriptions, and the websites of sword collectors, enthusiasts, historians, and armouries. Nevertheless, a combination of user warrant [22.214.171.124] and organizational warrant [126.96.36.199] has also been considered, especially where it was preferable for developing a usable organizational resource. LAT users are limited to employees of the organization, so in this case these two forms of warrant were very closely related.
*Pre-coordinate headings vs. post-coordinate retrieval
The LAT employs both single- and multiword terms to refer to indexable concepts. However, the structure of the thesaurus and the form of the terms is intended to facilitate post-coordinate retrieval. Users would be able to combine terms in a keyword search using Boolean operators, truncation, and nearest-match suggestions. This would be particularly useful when users are looking for a term in a form they may not be able to remember offhand. One such example is “stock removal,” which may be difficult to recall exactly but could be located with a search for “stock*” or “stock removing” under the system outlined above. Where used, compound terms have been constructed according to the parameters of the ANSI/NISO standard [7.1-7.6].