Mill Marginalia Online aspires to digitize all pages within the John Stuart Mill Library containing any mark or annotation while offering users a chance to view these individual pages within their larger textual contexts by providing links to already-available full-text versions of the marked books. In its original version, the site employed a hybrid of a SQL and noSQL database, one that capitalized upon the former’s speed in filtering and querying and the latter’s flexibility and ease of updating. Thus, a simple differentiation between verbal and nonverbal forms of marginalia was married to a Json field containing a readily expandable set of categories and subcategories that both allowed for more precise identification of the myriad types of marginalia found in the Mill Library, and accommodated the discovery of new forms of marginalia as the data was collected. This hybrid approach permitted such additions without requiring the consequent updating to the pre-existing dataset and large-scale data migrations that would otherwise be necessary in an SQL-only database. This custom-coded approach remained in place for roughly five years, including two years of development and three years of public use, at which point roughly one-fourth of the books in the Mill Library had been surveyed and their marginalia digitized.
In 2021, the site was entirely rebuilt using the PHP-based open source content-management system, WordPress. A new backend was created by the College of Arts and Sciences Office of Educational Technology and a new frontend was designed by Code Pixelz Media. The database shifted from the earlier hybrid to a MySQL format. The redesign preserves the functions of the original site while greatly simplifying the process of uploading new data and maintaining the security and multi-platform and browser compatibility of Mill Marginalia Online.
By 2023, the sheer amount of data collected (roughly 31,000 individual examples of marginalia and over 20,000 images) had begun to test the technical limits of the redesign. With the technical assistance of Shelbybark, LLC, and to preserve the stability of the site, a more limited WordPress interface was introduced. The project is currently seeking external funding for a third and hopefully final iteration of the site that will be able to accommodate the expected 20,000 additional examples of marginalia and their associated photos.
Behind the scenes, the project continues to employ a non-relational, relatively simple, and comparatively flat metadata schema that assigns marginalia the role of primary artifact, rather than classifying it as the secondary attribute of the book in which it appears. This self-conscious reversal of the hierarchies that underlie both TEI and conventional library cataloguing foregrounds the moment of human:book interaction and highlights that unique portion of Somerville’s Mill Library, the verbal and nonverbal marginalia, which holds the most promise to generate new knowledge. The project schema, too, is easily updated as new forms of marginalia are discovered.
Together, the database and metadata schema allow for multi-faceted searching of all verbal and nonverbal marks, which means that scholars interested in both Mills’ readerly judgments, writerly influences, and intellectual networks can consult and compare not just their annotations, but their broader strategies of marking their books. Even scholars without a singular investment in John Stuart or James Mill, but instead a diverse set of interests in cognitive approaches to textual studies, histories of reading practices, and literary aesthetics should find empirical evidence to support their research. Knowledge of the nineteenth century is literally inscribed in the margins of the Mill Library, and Mill Marginalia Online seeks to make that knowledge as accessible and broadly useful as possible.