Director, Electronic Text Center
University of Virginia
D-Lib Magazine, July/August 1997
Since opening as a full-time service in 1992, the Electronic Text Center
at the University of Virginia Library  has
pursued twin missions with equal seriousness of purpose:
Our work with collection management, archival imaging, and TEI-encoding has been reported elsewhere ; this article will focus on the integral place that our user community has in shaping the work of our library-based Etext Center.
Building a user community
The training of users is at first glance the more pedestrian part of what my staff and I do -- a task less exciting than adding a new etext to the online collections, setting up the latest Chadwyck-Healey database for on-line delivery, or improving our "on-the-fly" TEI-to-HTML filters. I am convinced, however, that our frequent and repetitive series of short courses in TEI, HTML, text-analysis software, and scanning techniques  are a central reason for the vitality of our humanities computing endeavors at Virginia.
Well-crafted training sessions, coupled with the ad hoc user support given to walk-in clients in our digital Centers , provides the "primordial soup" from which rises both a general literacy in the use of on-line resources, and also a growing suite of individual research and teaching projects. Between 1992 and 1997 we have built a considerable momentum, and more of our user support time now goes into the management of projects and a higher level of advice.
Users as a shaping force
Initially, we crafted our services with a firm sense of what our faculty and students would want, but with no fundamental input from them; even in 1997, humanities computing in the universities is too rarely driven by clearly articulated faculty demands. Now that we have created a growing number of users who are adroit in the use of our online collections, we are enjoying increased feedback on all aspects of our work: collection development, text encoding, image type and quality, and the design of the search and browse pages. A central lesson has been the validation of some of our primary choices: the 24-bit TIFF color images and the TEI texts that we have produced for years -- somewhat on the promise of future software and hardware developments -- are providing the flexibility we need to respond to a changing web environment and a developing set of user demands.
A stable, local user community, once created, becomes a great asset. It is an advantage that a university has with respect to a museum or a national library, where there are common types of users, but not necessarily the same individuals using the digital materials year after year, and growing more sophisticated in their needs and demands. The lessons learned over time from such long-term user feedback are now one of our greatest assets.
An active group of local users is salutary in other ways -- it prevents us from forgetting that we are a local service with a defined mission, even while we enjoy the fact that we have online users from dozens of different countries every day (circa 40,000 hits / 18,000 accesses / 4,500 unique hosts per day), and a high visibility in the national etext community. Indeed, much of our wider success is firmly predicated on our local focus, and the grounding effect this produces. The dozens of suggestions, queries, and comments that we get weekly from our non-UVa patrons are highly valued -- there is nothing like having thousands of users every day to shake out weak spots in a system -- but they do not take the place of the lessons one learns from training a faculty member in the use of etexts, or their creation, and then watching that person apply those skills over time in ways germane to his or her teaching and research.
The majority of our local use is by students or faculty whose needs are defined by the particular paper or book they are writing. However, a growing subset of users are making the move from being consumers to being producers of etexts and images. Their motivation for doing this can typically be categorized in one of the following ways.
1) Texts for study
For some of these advanced users, the best service we can offer is a partnership in the creation of an electronic text for inclusion into a course's reading list. Patricia Spacks, the chair of our English Department, was an inaugural adopter of our services. For the past four summers she has created an accurate computer file of a text that she would like to teach, but for which no available print text exists; we take the word-processing file, convert it to SGML, scan any images associated with the book, and add it to our collections. Our collaboration with Pat Spacks has added the following titles to our Modern English collection:
The division of labor that these titles represent allows us to "project manage" a large number of text-creation projects with a modest number of staff. Moreover, the contributing faculty or student feels some involvement in and ownership of the final product, and the process is de-mystified for them.
2) Texts as study
In addition to texts created by a faculty member in order to allow his or her students to study them, a group of our texts have been created by students as part of their classroom work. In a number of his courses, Jerome McGann has encouraged groups of his students to build electronic editions as a way to examine a text and to put into practice editorial theory learned in the classroom. Several examples can be found as part of our British Poetry Archive -- http://etext.lib.virginia.edu/britpo.html -- including Richard Polwhele's The Unsex'd Females (1798) and Mary Robinson's Sappho and Phaon (1796).
Other students have discovered that the web provides a medium to undertake a form of graduate paper not achieved with any satisfaction on paper alone, and our faculty have been willing and level-headed about allowing such essays. One example among many is The Ladies: A Journal of the Court, Fashion and Society (1872) by Virginia H. Cope, University of Virginia English Department: http://etext.virginia.edu/ladies/ladyhome.html.
By being both essay and facsimile publication, this examination of a rare nineteenth-century journal "offering scientifically precise fashion advice and demanding political rights for women" exploits the ability of the web to make available a hard-to-find publication as a part of an examination of that work.
3) Texts as research
Not surprisingly, the large searchable text and image archives at the
Electronic Text Center support a myriad of research uses, from high-schoolers
looking at 19th-century African-American manuscripts for a paper,
to faculty writing articles and books. A couple of examples will suffice:
4) Texts as archives
Sometimes, the creators of our texts come from within our library systems, as is the case with Kendon Stubbs and Sachie Noguchi, the editors of the University of Virginia/University of Pittsburgh Japanese Text Initiative: http://etext.lib.virginia.edu/japanese/
More and more, we are creating digital archives of local history, special collections, and other archival materials in our care, driven by the twin stimuli of preservation and access. Except in the act of selection, what a library creates tends to be -- appropriately enough -- an archive presented with no distinct point of view and for no one single audience. What we now see, however, is our users excerpting from and shaping these digital archives into much more focused and purposeful productions, often as "textbooks" for the classroom.
5) Texts as textbooks
For the past two years, Stephen Railton (UVa English Department) has worked closely with the Electronic Text Center and Special Collections to craft a site -- Mark Twain in His Times -- whose purpose is to provide a series of contexts around the Twain novels he teaches to undergraduates and graduate students: http://etext.virginia.edu/railton/. The site makes extraordinary use of the power of color digital facsimiles of manuscript and typescript items to invest these "great books" with some sense of their place as literary, economic, and social events. Contemporary reviews rub shoulders with images of Twain manuscripts; searchable texts and book illustrations exist alongside hypertext maps and sound files.
The content and the presentation of the site has been significantly informed by what Steve Railton has learned from his students over several semesters of use, and two of the loudest lessons have been the high premium they put on graphical content and on searchable texts -- they expect visual evidence and they want to interrogate and not simply to browse. The site is also testament to how much a partnership between a library and a faculty member can produce, at this point without any outside funding at all.
Effective teaching tools do not need to be wide-ranging and multi-semester productions to be impressive. If success is measured by the degree of utility to the users, then a technically simple site such as "Augustus: Images of Power" (Mark Morford, Classics Department) scores highly: http://etext.lib.virginia.edu/users/morford/augimage.html. This "slide lecture on a web page" makes the difference between these images being seen once in the classroom and them being a readily-recalled resource for the entire semester. A single HTML training session of the sort we routinely offer, coupled with a little follow-up support, was all the technical training necessary to facilitate this site.
6) Ancillary support
We learn time and again that a little support goes a long way. The regular provision of training courses, the public availability of the scanners and other machines in the Electronic Text Center, and the fact that there is always a staff member there ready to help, is enough to support a large body of users who are mostly -- and sometimes quickly -- self-sufficient. The difference between a small amount of help, and none, is the difference in many cases between these projects existing, or not. Below are three examples of recent users whose impact on our time is modest, and for whom we provide disk space, access to equipment, or occasional HTML and imaging support:
Users as a Resource for Future Planning
If our users get their money's worth out of us, as we hope they do, we increasingly are able to draw on them as a basic resource. This is nowhere more true than in the planning and shaping of a major text and image archive such as Early American Fiction (1775-1850): http://etext.lib.virginia.edu/eaf/
We are currently in the process of creating full color image and SGML text versions of 558 volumes of early American fiction, drawn from our Special Collections holdings. This major undertaking, funded by the Mellon Foundation and the University, has the analysis of usage patterns as a significant goal. To achieve this, we will make good use of the fact that we have local American Studies students and faculty for whom searchable archives are nothing new, and who therefore have some real experience and perspective on the issues we will ask them to consider. Even while we create the material we will be asking them to put it to use in their research and teaching, and expect to learn significant lessons about the eventual shape(s) of the Archive before it becomes available to a wider audience. In this, as in our evolving work with the web interfaces and "on-the-fly" HTML filters for our commercial and publicly-accessible text databases, the input of a large and rapidly maturing user community is invaluable.
While repetitive "user-training" in its most basic form may not be the most stimulating part of a digital librarian's work, it is a fundamental element in the task of bulding a user community for electronic resources. The building of such a community is in turn a fundamental necessity in a vibrant digital library. The provision of electronic materials is not enough in itself, now any more than it was five years ago when we began our efforts. It is not simply that supporting a user community gets the most value from the data one buys, and garners support for what is sometimes a new and somewhat risky undertaking within an institution. More important, and more exciting than that, the community you create quickly becomes a vital part of the resources on which you draw; users create texts, re-purpose existing product, and provide a high degree of critical observation on the various aspects of the digital library in which they invest themselves.
While the rate of usage is higher for those months that school is in session (see April's figures), and predictably dips on major holidays (see July 4th), the relentless rise in traffic month after month means that by July 8th the numbers are close to what they were at the height of the previous semester:
April 08 1997: high school and colleges still in session -- current average "school year" daily usage:
38,538 hits / 17,199 accesses / 4,606 unique hosts
July 04 1997: major US holiday: US traffic drops but not the rest of the world:
16,712 hits / 7,906 accesses / 1,769 unique hosts
July 08 1997: colleges in summer session only; despite this, numbers are heading towards last semester's usage stats:
33,605 hits / 13,519 accesses / 2,855 unique hosts