Static Search: An Archivable and Sustainable Search Engine for the Digital Humanities

Abstract:The end-goal of the Endings Project—a collaboration between project leaders, programmers, and librarians to address long term sustainability of digital humanities resources—is to have completely static sites: websites composed of only HTML, CSS, and Javascript that have no reliance on server-side processing. Now in the last phase of the grant cycle, the Endings Project successfully converted its major projects—like The Map of Early Modern London, The Robert Graves Diaries, and other HCMC-hosted projects—into static sites, but one final problem remained: search. Most search engines (like Solr) require the use of a server to index content and while Javascript search engines, such as Lunr, do exist, they cannot feasibly handle the vast amounts of data that comprise the standard digital edition.

This presentation outlines the creation of Static Search: an open-access codebase for creating a completely client-side search engine for static websites. Built using XSLT3, Saxon, and Ant, Static Search creates a JSON file for every distinct stem in a document collection and harvests metadata from each document containing that stem, providing a rapid mechanism for querying a document collection. In its current version, Static Search provides boolean searches (CAN CONTAIN, CANNOT CONTAIN, MUST CONTAIN) as well as exact phrase searching alongside faceted search filters based on document metadata. While currently implemented only for modern English, we are developing methods for querying early modern English as well as early modern and contemporary French as well as adding mechanisms for wildcard searches. Our presentation thus demonstrates both the feasibility of creating a static site with robust search capabilities and the advantages of Static Search for complex digital humanities projects.