In January, a working group of members from the Society for Classical Studies (then the American Philological Association), the Medieval Academy of America, and the Renaissance Society of America submitted a proposal to the Andrew W. Mellon Foundation for a planning grant "to determine the feasibility of creating a comprehensive working instrument for students of Latin, from the beginnings to the Renaissance, tentatively called the Digital Latin Library (DLL)."
In May, the AWMF notified us that we had received the planning grant. We had our first full meeting of the working group in July in Philadelphia. The second meeting was in December in Norman, OK. The third meeting was in May 2013 in Philadelphia. In between those meetings, a smaller group of scholars met in Norman to discuss plans for implementing the various parts of the DLL.
By the end of the planning period, the working group decided that the DLL project was worth pursuing. They appointed Samuel J. Huskey as the director of the project, and they established the University of Oklahoma, where Huskey has a faculty appointment, as the project's institutional home.
In January, the OU DLL team submitted a proposal to the AWMF for for the initial year of a projected three-year path to full implementation of all components of the DLL, "a Linked Open Data (LOD) online resource for finding, reading, discussing, studying, teaching, editing, annotating, and publishing Latin texts of all eras." The AWMF funded the proposal and work began in July. In addition to co-PI's Huskey, Abbas, and Weaver at OU, the DLL team included several graduate and undergraduate research assistants and three subcontractors: Hugh Cayless (Duke Collaboratory for Classical Computing), Tom Elliott (NYU-ISAW), and Alex Ward (independent IT developer).
During this funding period, the DLL team produced a user behavior study to collect information on how potential users of the DLL use resources that are currently available on the Internet. June Abbas conducted detailed interviews of several different types of users so that we would have information on what people do and do not like about the resources they use. With help from a graduate assistant, June organized and analyzed the results and produced a set of data for use in the implementation phase.
To aid in the development of the data model for the DLL's critical editions, Huskey and a team of research assistants reverse-engineered an existing critical edition and encoding it in XML according to the standards of the Text Encoding Initiative. The edition selected for this part of the project was Giarratano's edition of Calpurnius Siculus' Bucolica, since it is short (seven poems), has a full critical apparatus, and includes practically everything that a text editor might encounter while compiling a critical edition (line transpositions, corrections, a complex manuscript tradition, etc.). We will use this edition to test the various components of the DLL over the next two years of the project.
Chris Weaver and his graduate research assistants refactored the libraries for his Improvise visualization application in preparation for development of a series of visualization tools for use with DLL editions.
Abbas, Huskey, and Weaver evaluated a number of potential applications for the DLL's catalog, bearing in mind data from the user studies.
Several experts in the field of textual criticism came to Norman to record videos for the DLL's video library.
The DLL convened several scholars for a seminar on Latin textual criticism in the digital age.
In January of 2015, the OU DLL team submitted a proposal to the AWMF for a two-year implementation grant to build and implement the Digital Latin Library (DLL).
The original plan was to deliver the following collection of resources collectively known as the Digital Latin Library (digitallatin.org) by the middle of 2017:
As that deadline approached, it became clear that an extension would be necessary, owing to the complexity of the project. In June of 2017, the AWMF granted an extension of one year. The current goal is to launch all components by mid-2018.
Work on the DLL’s catalog continues apace. Models for the data and metadata have been established, and the information architecture is in place. The catalog runs on Drupal 7 with a MySQL database and Solr search engine. Currently, there are 2,377 individual author authority records and 4,236 work records, spanning the Classical, Medieval, and Neolatin eras. With that framework in place, Huskey has been using various data scraping techniques to build out the catalog’s item records with content from a variety of sources including the Packard Humanities Institute’s Latin Texts, digilibLT, Biblioteca Italiana, and more.
Huskey and Cayless have established encoding guidelines for the LDLT. A pre-release version is available at https://digitallatin.github.io/guidelines/LDLT-Guidelines.html. This work has enabled the other parts of the project to proceed.
Cynthia Damon, Robert Kaster, Andrew Dunning, and Jeffrey Witt are currently working on pilot editions that will be used to test the guidelines and the other components of the project. Huskey has been working closely with the editors of these pilot editions to identify ways to facilitate the tedious work of encoding texts. His undergraduate student Virginia K. Felkner has been developing his initial work on automating encoding processes with Python scripts. These scripts can take plain text files of an edition's text and CSV files of apparatus criticus data as input and produce well-formed, valid TEI XML in accordance with the DLL's encoding guidelines. By the time of the project’s official launch in mid-2018, it should be possible to limit the amount of encoding work to the handful of cases that will be unique to any text.
The three learned societies affiliated with the DLL project are developing policies and procedures for peer-review and publication of editions for the LDLT. The goal is to have processes in place to begin accepting proposals by the end of 2018.
Weaver and his graduate students have developed a number of visualizations for LDLT data using Weaver’s Improvise visualization system. These visualizations include one that resembles a traditional print edition, a story-line based visualization that tracks witnesses entry by entry, and a pixel-based visualization that brings together multiple data points into one interface. These visualization techniques demonstrate what can be done with Latin texts encoded as machine-actionable data. Weaver is currently finalizing the visualization application for the LDLT, with the goal of releasing it by mid-2018.
Finally, under the direction of Tara Carlisle, Digital Scholarship Specialist at OU, a team of researchers is currently working on a search application for openly available Latin texts. This web-based application uses ElasticSearch to index texts that have been cataloged and added to a database. The initial launch will feature texts from http://thelatinlibrary.com, but texts from other sources will be added over time, with the goal of providing a comprehensive search engine for Latin texts of all eras.