Submersion in the question of keywords
For more than 20 years, the online Pallas catalogue has allowed to search documents in the collections of CegeSoma. Pallas proposes different ways of navigating to the users, among others the search by keywords.
From the results of the MADDLAIN project, we know that the success of a search by keywords is directly dependent on the way in which the documents have been indexed. For example, if all the keywords have not been translated in a systematic way, a French-speaking user and a Dutch-speaking user entering their request in their respective languages will not obtain the same documents in the results list. Moreover, if many similar keywords have been created during the cataloguing, the user may be bewildered when finding that his query for the words “Belgian resistance” results in dozens of suggestions of longer keywords, containing his two original words but also other, more precise, terms.
These examples take us to the question of the quality of the metadata. In the context of the research projects ADOCHS and UGESCO, an analysis of keywords used by CegeSoma has been initiated with the aim of improving the digital access to the collections. The goal is to find short-term solutions for optimizing the search results, but also to conduct an in-depth reflection on the way in which the quality of the keywords can be controlled and improved within the institution. For instance, the short-term solutions include the translation of the 100 most frequently used keywords that have not yet been translated in Dutch or French. The ADOCHS project also examines how the development of the linked data can, in the longer term, allow to improve the quality of the keywords, particularly by using databases such as Wikidata, GeoNames or VIAF.
By focusing on the question of keywords, CegeSoma wishes to respond to a clearly expressed need by the users with regard to the improvement of access to the collections.