Tag-Based Browsing of Digital Collections with Inverted Indexes and Browsing Cache

In this paper we describe one of the browsing strategies for learning object repositories implemented in the Clavy platform, a platform for the management of repositories with reconfigurable structures. Since Clavy makes it possible to dynamically modify the structure of the repositories, it must support a navigation model independent of that structure. For this purpose, the platform adopts a tag-based browsing model, according to which users select descriptive tags to filter learning objects (in Clavy, these descriptive tags correspond to element-value pairs). In our experience using Clavy, we have realized that updating the browsing state when the user changes the set of selected tags can be a costly process. The proposed strategy alleviates this cost by combining inverted indexes with a multilevel cache model that enables the system, on the one hand, to cache filtered objects (i.e., set of objects filtered by sets of selected tags), and, on the other hand, selectable tags (i.e., set of tags able to shrink sets of objects).