The digital collection “Poe’s Magazines” will include scanned images, structured texts, associated metadata, authoritative issue previews, and biographical sketches of contributors. 95 issues from the four periodicals Poe edited will be processed for this project. Totaling approximately 3200 pages, the content includes 50 monthly issues from the Southern Literary Messenger, Burton’s Gentleman’s Magazine, and Graham’s Magazine; and 45 weekly issues from the Broadway Journal).
Digitization of the page images is guided by the standards published in the National Information Standards Organization (NISO) publication, A Framework of Guidance for Building Good Digital Collections (2007; http://framework.niso.org/.) Scanning by the LSU Libraries Digital Services provides a high-quality 600ppi image for each page. In order to replicate the appearance of the physical page as closely as possible, image manipulation such as color correction, lightness/darkness, or sharpening is generally not performed; however, the images are cropped to remove black scanning space and maintain the integrity of the physical page edge.
“Poe’s Magazines” follows the TEI: P5 Guidelines for Electronic Text Encoding and Interchange. Text representations are being created for each scanned page using the TEI XML format. The project team encodes basic structural markup to TEI Level 3, and level 4 elements that mark people, places, and entities. (For a description of TEI levels, see Best Practices for TEI in Libraries (version 3.0); http://www.tei-c.org/SIG/Libraries/teiinlibraries/).
There are an estimated 250 named authors (identified and pseudonymous) represented in the Poe-edited issues from the four periodicals. Local authority records for each identifiable author/editor are created according to the rules prescribed in Describing Archives: A Content Standard (DACS; http://files.archivists.org/pubs/DACS2E-2013.pdf) using the Encoded Archival Context-Corporate bodies, Persons and Families (EAC-CPF) 2010 XML schema (http://eac.staatsbibliothek-berlin.de/). These authority records link name variations and pseudonyms between publications allowing for inter-collection consistency and potential external interchange with other repositories. In addition, these records offer information on the biography, history, and relationships between contributors.
The Islandora Project open-source software framework serves as the primary platform for repository management and public access. The project uses the Islandora Project Book Content Model to encode each magazine issue. Islandora generates Dublin Core and Metadata Object Description Schema (MODS) records for each periodical issue in the repository. The repository stores each page’s original high definition image and generates derivative (PDF, JPEG and JPEG2000) images. Plain text is inserted into each page record as it becomes available, and TEI representations are authored using the Islandora toolset.
Additional content for the four periodicals includes authoritative previews of each issue and the searchable EAC-CPF records of contributors including biographical sketches. As part of this preview, an overview of each issue and a table of contents encoded via TEI is also created. Themes for each issue are currently being marked subjectively issue-by-issue, but as the process continues we will be experimenting with adding computationally derived themes using topic modeling. The EAC-CPF records are created using customized Islandora functionality and stored as an auxiliary collection within Fedora. Where authoritative records are available, we make reference to them and in all cases, we use the record as a container for local contextual information. Where authors’ names are not known, we assign a unique identifier based on the issue, date, page number and article title. Identification of anonymous authors will be a potential topic of future crowdsourcing research.