Has the tree structure outgrown its usefulness or is the “tree of knowledge” image so powerful that it remains a fundamental principle of knowledge organization? This was a recurrent question first raised by Scott Weingart of the University of Indiana in his paper “From Trees to Webs”. It remained a prevailing theme of the last biennial conference of the UDC Consortium, which I attended in October last year in The Hague.
The title of this international seminar was “Classification and Visualization” and it provided an excellent opportunity to discuss and reflect upon the interactive, web based visualization tool we have included in our new beta ELSST application. Such interactive applications can respond to the demands of web 3.0 users for fast, intelligent browse options but will also safeguard the integrity of the live source in a way that tree structures alone do not.
At this stage of our project, while development work continues and our work on inconsistencies between the two thesauri progresses, our focus is upon thesaurus structure.
Some of the structural issues we have been addressing throughout the inconsistency work are:
· Hierarchical structure (a mono or polyhierarchical structure)
· The optimum number of top terms
· How to ensure currency and avoid redundancy of terms
· The axioms and constraints for aligning two thesauri with sufficient inbuilt flexibility for editing requirements
At the CESSDA-ELSST meeting of 31st July 2013 Lucy Bell raised the issue of polyhierarchies and facet analysis and invited all to contribute their ideas. There has been some very helpful feedback regarding facets and classification and preferences for browsing functions expressed.
One comment made at the ISKO UK Conference 2013 was that as few as a dozen top terms might be sufficient for the social sciences. This would be a subject based structure. Another suggestion made was to consider the model of the GESIS Thesaurus, the Soz, which identifies a category of ‘special social sciences’, as well as one labelled ‘interdisciplinary application areas of social sciences’, within a browsing structure with six major categories.
The message from the papers and discussion at the International UDC Seminar 2013 was clearly that we can utilize new technologies to solve some of the problems that an emerging ‘big data’ environment poses for knowledge organization. Paul Otlet’s review of the Dewey Decimal Classification at the end of the nineteenth century, and the subsequent development of the Universal Decimal Classification, was presented as an example. It is currently being used in 130 countries and published in over 40 languages. It is also available in SKOS format. We are exploring a possible bibliographic application and will post further on this work.