Principal Investigators: Joseph Loewenstein, Washington University in St. Louis; Patrick Cheney, Pennsylvania State University; Elizabeth Fowler, University of Virginia; David Lee Miller, University of South Carolina; Andrew Zurcher, Queens' College, Cambridge
The Spenser Archive is the digital component of Oxford University Press's forthcoming Collected Works of Edmund Spenser. A multi-institutional undertaking, the Spenser Archive provides scanned images of original editions, edited texts with in-depth markup, editorial commentary, text and scans of variants, with a flexible display of Edmund Spenser's works.
Spenser Project Press
- The Paragon project, an NEH-funded project at South Carolina to build a digital collation tool.
- UVa Today piece on Elizabeth Fowler and NEH grant (includes video of Fowler reading from Faerie Queene Book I) (14 December 2007)
- WUSTL's Record piece on NEH grant (14 October 2007)
- "Building a Spenser Archive — One Scan at a Time" by David Miller (Duke University Libraries, Spring/Summer 2007)
Core Data Management Functions
The HDW delivers core data management functions to the Spenser Archive. These functions include:
- The collation database, a central, secure web facility where the project maintains information about Spenser's works, the editions of those works, witnesses to those editions, and differences between the witnesses. These data elements meet the core information needs of the textual editing process. We currently have record of about 3,600 differences and other observations. In addition, we use the collation database to maintain a finding list that contains information about ~1500 copies of early editions of Spenser's work held across an international set of ~260 libraries. Lastly, we provide editorial tools (facilities for viewing page scans, readable versions of the texts we're preparing, and tools for analyzing textual stemma and forme states). The collation database is built upon the HDW's standard, open-source relational database management solution. Data is stored in a MySQL database. Users access the database through an HDW-customized version of the Django admin functions., and additional Django customizations were written specifically for the Spenser Archive.
- An infrastructure for marking up texts in TEI. The Text Encoding Initiative (TEI) is an international working group dedicated to the development of a standard for encoding electronic texts. TEI is a variant of XML, and contains the full set of features necessary to capture all of the information relevant to the preparation of a scholarly edition. The HDW provides a number of services and facilities in connection to TEI:
- Expertise in TEI, and in maintaining TEI-compliant schemas.
- Data management. The project's XML files are stored on and served from a centrally located SVN repository. SVN, most often used by software development project teams, suits our project by providing centralized storage, backup, versioning, and update control.
- Tools for manipulating TEI documents. For most changes, the project uses Oxygen, a commercial XML editor; however, the HDW also provides a custom web-based XML editor, suitable for small changes. We also provide stylesheets for transforming TEI documents into other formats, usually formats suitable for human readers.
We understand Spenser's works as belonging to a dense network of other texts, and when the Spenser Archive becomes publicly accessible, we are going to want to represent at least a portion of that network. However, in implementing those relationships, we face a number of technical limitations most of which originate in the rather meager capabilities of the standard HTML anchor tag. To try to work around these limitations, we devised the a new set of tools for linking TEI documents, and for displaying the results of the those links. We wanted the new set of tools to:
- Maintain physical separation between texts. TEI documents containing Spenser's works should contain just Spenser's works, and not any other texts, or even any links to any other texts.
- Support bi-directional links. The standard HTML anchor tag goes only one way, from one text to another, which has the effect of privileging one text (the originator of the link) over others (texts linked to), and of forcing the reader to read in a cerain order (read the privileged text, then the texts it links to).
- Permit editors to write notes and commentary using whatever tool they're comfortable with. Once editors have written commentary, we convert it to TEI, then link the commentary TEI to Spenser's texts.
- Categorize links, so readers can select one kind of link or another, as they desire.
We've used these tools to link Spenser's texts and editorial comments, and to prepare sample texts for the Oxford University Press. We're also planning on using them for the Life of Apollonius project.
During the fall of 2009, students at Washington University and Pennsylvania State University undertook an inter-editional collation of the five editions of The Shepheardes Calendar published Spenser's lifetime. The work involved comparing four editions, published in 1581, 1586, 1591 and 1597, against the first edition, published in 1579. Texts were compared line-by-line, and even the smallest differences were recorded. In all, the students recorded over 13,000 differences between the five editions.
Given the amount of data generated by the comparison and the need to make use of it in a manner commensurate with the effort undertaken to collect it, the HDW felt that a specific set of tools were needed to organize the data and facilitate analysis of it.
Teaching and the Project
Much of the work on the Spenser Archive at Washington University happens in the classroom. Professor Loewenstein typically meets with students for an hour or so each week throughout the semester. While some of the meeting time addresses the mechanics of organizing the project (how are tasks progressing? who is doing what next?), much of the time is also directed toward whatever questions come up. Usually those questions are about the immediate matters of textual criticism and physical bibliography, but they can also be about broader issues of the historical and cultural contexts of Spenser's work.
Behind these meetings and questions is an analogy of the project as a science laboratory. In a laboratory in the sciences, we might expect that a professor leads the research, and that he or she employs students as apprentice researchers. In effect, professors do more than research--they also train the next generation of researchers. Likewise, students don't just work, they also learn acquire the skills to grow into increasingly responsible research roles.
This model has served the Spenser project, the HDW and its students well, and is one that the HDW seeks to reproduce in other projects.
Several years ago, we had an undergraduate on the project named Megan, who took the first stab at setting the text for an edition of A Theatre for Worldlings (1579), a anti-Catholic screed to which Spenser contributed translations from Petrarch and Marot. Megan, working by hand through the collation data, organized our observations of the differences between various witnesses of the edition, and was thus able to specify how the edition had been corrected as it was printed.
Robo-Megan started two traditions with the project and with the HDW in general:
- While a standard toolset like our Django platform can go far toward meeting a project's needs, every project would seem to have unique requirements. Where possible, we like to build software for those requirements around our standard Django platform, as we have done here.
- A surprising number of our best ideas come from our undergraduates, so we pay attention to what they're doing, since their project-related activities will often surface an opportunity for automation. And, in recognition of that fact, we like to name software after them.
Megan graduated, and has gone on to work for a non-profit in D.C. But we're reminded of her contributions to the Spenser Archive every time we run Robo-Megan!