Lyle H. Wright’s American Fiction 1851‐1875, A Contribution Toward a Bibliography (San Marino: The Huntington Library, 1957) has long been recognized as the standard bibliographic compilation of mid nineteenth‐century adult novels, novellas, short stories, and other works of prose fiction published during that 25 odd year span in the USA. While many of the titles included in this source are major canonical works of American literature such as Moby‐Dick and The Scarlet Letter, the vast majority of the works included are rare and often inaccessible to most researchers. At a time when the study of non‐canonical works is very much in fashion, finding copies of works by long forgotten authors becomes paramount for literary scholars. The bibliography does contain a “census” of the holdings of major libraries such as New York Public and the Library of Congress, but such information is of small consolation to researchers who have been forced to use microfilm sets that may be equally unavailable to users from smaller institutions. Thus, the works cited by Wright are prime candidates for digitization. The Committee on Institutional Cooperation (CIC), a consortium of Big Ten universities and the University of Chicago, and the Digital Library Collection of Indiana University, which serves as the hosting site, has stepped in to provide a searchable online collection of the famous as well as the obscure texts in that compilation. The Wright American Fiction Project attempts to digitize all the works cited in the bibliography and to provide an online library that will fill the gaps in coverage of American fiction found in other digitization initiatives. Currently 1,932 texts of the 2,823 in Wright’s list have been included, though only 201 have been fully edited and marked up with SGML encoding. The ongoing project will give researchers unprecedented access to nineteenth‐century literary texts when it is completed, and it now offers a good example of how a well‐organized digitization project can produce an invaluable resource for all levels of literary study.
The collection follows strict standards and principles in its digitization practices and represents a collaborative effort on the part of many participating libraries. Nine of the CIC libraries committed staff and resources to the first phase of creating digital images from the microfilm photographs of the text. Each image consists of a single microfilm frame displaying two pages of print. The images are stored offline in TIFF format, and five CIC institutions oversaw the conversion of the images into electronic format using Optical Character Recognition (OCR) software. SGML encoding allows the developers to make the texts fully searchable and preserve structural features such as Tables of Contents and other divisions. Although errors in the OCR processing do occur, the end result is a faithful rendering of the texts, complete with typographical errors and idiosyncrasies of spelling or printing. Researchers viewing the works in the collection can see the pages either as plain text or as GIF images, and Indiana University has acquired and modified software from the University of Michigan to create a search engine and their own unique display interface. The resulting database offers many different search options and many options for viewing texts and creating personalized “bookbags.”
Among the search features in the Wright American Fiction Collection are a browsable list of authors (with the number of titles currently available for each) and a word index of all unique words contained in the texts. Simple searching mode allows users to enter words or phrases to be searched in the full text, chapters, or paragraph headings of works. More advanced search options enable users to do proximity searching (“near,” “not near,” “followed by,” “not followed by”) and to use Boolean operators to search for terms within chapters and paragraphs. The database also has standard citation searching to locate works by author, title, publisher, place of publication, and year of publication. The search history option is a useful tool for more advanced users doing multiple searches. Retrieved records can be stored in a bookbag or online shopping cart, and users can do simple, Boolean, or proximity searches within the titles they have selected. This list, unfortunately, cannot be saved once the session is closed, but users can e‐mail the records to themselves. Records for fictional works in the database have hyperlinked chapter headings for navigation and search options within that text. Users can also choose to download the entire work into their browser (the OCR‐created text, not the GIF page images) if they wish to save it. The “view entire text” link takes users to a screen with instructions as well as a warning about the size of the download. Within each chapter of a work, it is possible to view scalable images of the two‐page microfilm frames or the plain text, and navigation tools are readily accessible.
Indeed, compared to many digital text sites and databases currently available on the Web, the Wright collection is rather intuitive in its design and relatively easy to navigate. Some novice users seeking to perform basic tasks such as saving the texts encounter some difficulties, but the help screens contain a good set of tips and tricks for using the database, and the FAQ list covers many of the basic questions. The link to the help pages appears only after users have selected a particular search mode, so having a more prominent link to that valuable section of the site on the home page would be a good future enhancement. The Wright American Fiction Collection represents a great advancement in the application of digitization technology to preserving printed texts and making them available to the widest possible audience. This site truly benefits both academics and even the reading public alike, and once the project is completed it will open up many new possibilities for research into a pivotal time in American literary history.
