We assessed wood-technology-related contents in a historic newspaper Kmetijske in rokodelske novice (KRN), digitized through the Digital Library of Slovenia (dLib.si), developed and supported by the National and University Library, Ljubljana (NUK). Prior to our additions to the KRN, no systematic subject headings were available for retrieval of wood-related contents. Based on the samples issued in 1861, we visually (manually) selected articles with wood-related contents and compared such visual selection with the results established on basic- and advanced-search principles in the same year, by employing different wood-related search terms. Digital searches were unsuccessful with the exception of the search term les, which resulted in a huge search-noise - results too numerous to be useful. Accuracy of retrieval with les in the digital version of KRN depended on the optical recognition of a century and a half old printed pages; so we also checked the reliability of the process. Optical identification of the word les in selected articles turned out to be sufficiently accurate and comprehensive, and could thus be used as a satisfactory tool for searching the entire corpus of the KRN. As experts in the wood science profession we then manually narrowed down 2098 articles - resulting from the searching for les - to 236 relevant articles, and classified them in one of the 16 more specific wood-related categories. We manually assigned 4 tags to each article: lesarstvo, lesar, les and the fourth tag (one out of the 16 content-specific tags). Total sum of the added tags was 944. We examined the possibility of adding tags in 5 most commonly used web browsers and 3 most commonly used operating systems in Slovenia. New tags enriched the subject headings of KRN. Future retrieval of wood-related contents in the dLib should consequently be more successful and more accurate. We suggest the inclusion of a wood-related subject category to the dLib topic viewer.
|
