Abstract
This article gives a history of the digital humanities project At the Circulating Library: A Database of Victorian Fiction, 1837–1901, from its beginnings in 2007 to the present. In particular, it recounts the mistakes and successes, joys, and frustrations of creating and maintaining the project, as well as some observations about working on digital humanities projects.
The Romantic era, through the work of Peter Garside, James Raven, Rainer Schöwerling, and Verena Ebbes, boasts a comprehensive bibliography of the period’s fiction through the book The English Novel 1770–1829: A Bibliographical Survey of Prose Fiction Published in the British Isles (2 volumes; 2000); its affiliated website, British Fiction 1800–1829: A Database of Production, Circulation and Reception (2004); and its appendix website, The English Novel, 1830–1836: A Bibliographical Survey of Fiction Published in the British Isles (2006).1 Across the three projects, the authors identify 1,421 fiction titles published from 1770 to 1799 (averaging 47.4 titles per year); 2,272 fiction titles published from 1800 to 1829 (averaging 75.7 titles per year); and 610 fiction titles published from 1830 to 1836 (averaging 87.1 titles per year). They found in total 4,303 works of fiction during this sixty-seven-year period. As their numbers suggest, the production of fiction increased remarkably from the late eighteenth century up to the cusp of Queen Victoria’s reign. To date, no such comprehensive bibliography exists for the Victorian period. Inspired by the efforts of Garside and his colleagues, the digital humanities project At the Circulating Library: A Database of Victorian Fiction, 1837–1901 (hereafter ATCL)2 aims to create a comprehensive bibliography of the Victorian novel. Currently, the database has entries for more than 24,000 titles, 5,500 authors, and 700 publishers; however, more titles and authors continue to be added every year, ultimately until every work of fiction has been identified.3 As these numbers suggest, ATCL has already documented eight times as many fiction titles in the sixty-five-year reign of Queen Victoria than Garside and colleagues found in the Romantic period—a fact that perhaps explains why they wisely stopped at 1836.4
The freely available ATCL has several ways for users to access the data on Victorian fiction publishing. The “browse” tab leads to lists of all titles, authors (with sub-lists of men and women authors), publishers, publication years, genres, fiction-carrying periodicals, and serializations (with sub-lists of novels in parts and magazine serials). The “search” tab allows users to search for individual titles or authors. The “analysis” tab offers links to pages of general statistics, data visualizations, and a random sampler (a personal favorite). The latter page also allows users to download the complete dataset as CSV (comma-separated values) files for use as they please. In the past two years, the website has been optimized for accessibility and to work across a range of devices from small iPhone screens to tablets to desktop monitors.
Previous entries in the Digital Deliverables section have focused on the particulars of their projects—details that I have found uniformly fascinating. However, for my entry, I would like to chronicle the history of our project, as best as I can recall it, for several reasons. First, I think it is important to preserve the labor history of large projects, especially digital humanities (DH) projects. The work we do often remains invisible to the larger public, so peeking behind the website may give insight into both the project itself and the field in general. Second, though I do not claim any great wisdom will be found in my narrative, I believe that reading about the mistakes and successes, joys and frustrations of creating and maintaining ATCL may serve as an inspiration (or possibly a warning) to aspirants in DH.
Now approaching its twentieth year, ATCL has grown significantly from its conception. Initially, based on a question from my dissertation (defended in 2002), I wondered exactly how many novels the Victorians produced in the three-volume format. Naively, I thought it would not be too hard to tally them up. So around 2004 I began to compile an Excel spreadsheet of three-volume novels in order to write an article that eventually appeared as “The Production of Three-Volume Novels, 1863–1897” in Papers of the Bibliographical Society of America.5 The titles were culled by going line by line through five volumes of The English Catalogue of Books, a relatively easy but time-consuming task (I recall doing some of this work sitting by a pool). When told of the list I had made, someone at a conference (regrettably, I forget who) suggested posting the list online, as it might be of interest to other scholars. A static list of titles thrown up on a webpage, for some reason, did not appeal to me, so I thought an online database such as the one Garside and his colleagues had made would be more useful. At the time, I was a contingent lecturer at the University of Kansas without access to funding, technical support, or the DH centers that exist today, so I simply went to the bookstore and bought a MySQL and PHP guide and taught myself just enough web programming to create a simple database and website.6 (I cannot emphasize enough how primitive the database and website designs were—an original flaw that would need to be addressed frequently in years to come.) As far as I can recall, I did not reach out to anyone knowledgeable in DH or web programming at the time; I simply did it all on my own.7 Around this time in early 2007, I accepted a job offer from my current institution, Purdue University Fort Wayne (PFW). Naturally, I asked their IT department to host my project, but after much foot-dragging and many circuitous email exchanges, it was clear they could not be bothered. At the Society for the History of Authorship, Reading and Publishing (SHARP) conference in Minneapolis that summer, I met for the first time Patrick Leary. Hearing of my project and travails during a chat at the opening wine reception, he on the spot offered to host my website himself, even jotting down the username and password to his server. I was stunned by his offer, and Patrick’s generosity that day and since still amazes me—though, to those who know Patrick, his enthusiastic support of scholars and their work comes as no surprise. To this day, I have never paid one penny to host ATCL.
Later that summer, the beta version of the website debuted online with a total of 1,700 titles, all three-volume novels. With the support of a PFW summer research grant in 2008, I added the final Victorian three-volume novel to the database by September 2008. Sending a message to the VICTORIA and SHARP-L listservs was the only publicizing of the website, done only at the persistent prompting of Patrick. I finally had the answer to the question I had posed back in 2005: The Victorians produced 5,168 three-volume novels, the final two appearing in 1897. I remain inordinately proud of this singular achievement, harboring a vision of this number eventually being engraved on my tombstone. In all honesty, I had intended to stop here, as my attention turned to writing enough articles for tenure and drafting a monograph on the end of the three-volume novel format. I had the answer to my initial question and more than enough data to write my intended book. Had I stopped, ATCL would likely have shared the fate of so many DH projects begun with bright-eyed enthusiasm—suffering a neglected, flickering half-life online until software obsolescence or turning off the server erased it from our digital memory.
Fortunately, two things kept ATCL growing after 2009. First, I had the good fortune to receive an National Endowment for the Humanities (NEH) Summer Stipend for a project to add Victorian two- and four-volume novels to the database. In combing through The English Catalogue, I had flagged other multivolume novels, and it seemed silly not to add these titles as well. So, by October 2010, I had added another 2,144 two-volume novels and 19 four-volume novels to the database for a total of just over 7,000 titles with corresponding entries for more than 2,400 authors and 200 publishers. Again, I had intended to stop here—a list of all Victorian multivolume fiction seemed a good and respectable stopping point. Not to brag, but ATCL at this point had already nearly doubled the size of the Romantic fiction bibliographies.
However, a question raised by my monograph research prompted the first major reconception of ATCL: How many of these multivolume novels had been serialized before publication? Therein lay a problem—the database had been designed to collect title and author information, not information about the dates and locations of serializations. At first, serialization information was simply added to the titles table, but that quickly became unworkable—hard to display on the website, equally hard to analyze for my research, and limited to one serialization per title. This moment, then, became the first big redesign of not just the database but the website as well. Without getting too technical, the database expanded from three basic tables (authors, publishers, titles) to five with the addition of periodicals and serials tables. Correspondingly, the website’s PHP code needed to be updated to accommodate these changes. For a good two years, I spent a substantial portion of my time tracking down serializations in Victorian magazines and newspapers (both digitally and at the British Library’s old Colindale outpost) and adding the information to the database, work funded in part by a Bibliographical Society grant (2010) and a Research Society for Victorian Periodicals (RSVP) Curran Fellowship (2014). Although I was not adding more titles and authors, I was steadily adding new information about the titles and authors that were already in the database.
By 2012 or so, ATCL had begun to attract greater attention from the public and other scholars, who generally rated it positively. Statistics from the server showed a steadily increasing number of users from around the world—amazingly, unlike a print publication, I can identify the number and location of my readers. I had begun to present at academic conferences, particularly those sponsored by SHARP, using the underlying data from ATCL or focusing on the website itself. This warm reception and the act of changing the database forced me to look at ATCL in a new light. It is hard to recapture my thinking at the time, but looking back, I began to see ATCL as more central to my research program instead of the peripheral project it had been when I started. The latter thinking had been reinforced by the fact that ATCL did not “count” as a publication for my tenure case, so whatever time I devoted to it was “wasted” time that could be better used writing articles and books that would count. This is not to say that my institution was hostile to DH in general or my project in particular—just the opposite, in fact—but that at the time there was no mechanism to evaluate it. My external letter writers for my tenure and promotion case that year all dedicated space in their letters to lauding ATCL as an already useful resource (sometimes at the expense of my print works).8 In a moment of hubris, I imagined ATCL becoming not just a comprehensive bibliography of the Victorian novel but a “key to all Victorian novels” à la Casaubon (a nickname my wife has jokingly used for me). Serialization information could be added, but there was still a need for more information about authors and publishers—a thousand of the authors in the database at that time had no birth or death dates and appeared in no other reference work. Going further, titles could be grouped by genres and linked to digital copies, authors could be grouped by nationality and occupation, and statistical overviews and graphs could be added. More important, the remaining Victorian novels published in one-volume editions could be included for the first time. The possibilities seemed endless.
I don’t recall the exact day when I decided to add one-volume novels to the database, but I do recall the first such title: William Makepeace Thackeray’s Vanity Fair. Working on the serializations of Thackeray’s other novels, it seemed priggish to exclude his greatest and best-known novel simply because it had the misfortune of appearing in a single volume. Quickly after, I likewise added Charles Dickens’s other fiction. From that point on, once that self-imposed restriction fell, all Victorian novels warranted inclusion and ATCL took on a larger life and purpose. Unlike the earlier bibliographical work, I followed no systematic plan but added new titles and authors as I found them, though I did work my way through John Sutherland’s The Stanford Companion to Victorian Fiction (1989).9 I began researching the lives of authors listed in the database that could not be found in existing reference works such as Sutherland or the Oxford Dictionary of National Biography. For this biographical work, digitized census records and newspapers have been a godsend, and I have been able to write nearly a thousand brief biographies of otherwise forgotten authors (if I am honest, my favorite kind of work). I also added database tables for genres (for titles) and groups (for authors)—now called “title tags” and “author tags” on the website. In some cases, genre-specific bibliographies proved helpful in identifying titles to add, for instance, Andrew Maunder’s list of sensation novels found in Varieties of Women’s Sensation Fiction: 1855–1890 (2004).10 Grouping authors together—as far as I know, the only attempt to do so in a systematic way—reveals some strange connections, such as the number of siblings who each wrote a Victorian novel, the number of clergyman-novelists, or the lesser number of governess-novelists.
Hence, from 2013 to the present, progress has been marked by the growing total number of titles in the database: over 10,000 titles by December 2013; over 12,000 titles by February 2016; over 14,000 titles by May 2017; and over 16,000 titles by June 2018. On this latter date, a notable tipping point was reached as the number of one-volume titles overtook the number of multivolume titles. Pivotally, the database served as the foundation for chapter 1 of my monograph, The Rise and Fall of the Victorian Three-Volume Novel (2020), a book that contributed to my promotion.11
Up until around 2018, I had worked alone; identifying novels, researching authors, data entry, writing code, website design, everything had been a solo effort. Except for the few research grants mentioned above and Patrick’s web hosting, I had received no research funding beyond what I had paid out of my own pocket.12 Without funding, I felt that I could not in good conscience ask students to assist me without pay. Good fortune again found me when Allen Riddell took a tenure-track position in the library and information science department at Indiana University Bloomington. When he was a graduate student at Duke University in 2011, Allen and I exchanged emails about the website. A few years later as a postdoc, he reached out to me again for help with a bibliometric paper that he was working on to estimate the number of novels published in the nineteenth century.13 After his move to Indiana, our correspondence increased, and our proximity allowed us to meet in person on occasion, with the drive between our campuses being about three hours. Allen brought new energy and ideas to the project that, frankly, I hadn’t known I was missing. First, he had ideas for using the data I had collected, so we began collaborations on creating a representative sample of fiction titles, studying fiction authorship in 1838, and analyzing digitization rates of Victorian novels.14 Second, Allen spearheaded our search for funding to accelerate data entry. Our many applications to the NEH and other programs all failed to secure funding—not from any serious failures on our part, but because ATCL seemed to fall between two stools. On the one hand, it was a “mature” project, and most grants are aimed at helping start projects, not sustain them. On the other hand, ATCL remains, let’s face it, a decidedly unsexy project. Many grants want to fund new innovations, such as artificial intelligence (AI) or machine-learning methodologies, and ATCL remains rooted in old-fashioned bibliography. Ultimately, we have become a victim of our success. Since ATCL succeeds now without funding, why would it need more money? I remain frustrated and disappointed by the dearth of research funding for projects such as ours in the United States. Allen did, however, bring additional institutional support and access to graduate students. One local research grant funded several student workers over two years to add titles from The English Catalogue; another grant paid for updating the website for use on mobile devices.
Much as adding Vanity Fair to the database had opened the floodgates to more novels, adding Allen to the project paved the way for more collaborators. I quickly identified a handful of individuals who had maintained contact with the project over the years and invited them to take on a larger role in ATCL, including Marie Léger-St-Jean and her DH project, Price One Penny, on penny serialized fiction15; Matthew Poland and his interest in magazine serializations; and Jessica Terekhov and her project on novels in parts. In short order, I went from a lone researcher to a team leader. They continue to challenge me to make the project stronger, already prompting a reconceptualization of the serialization modeling and a massive reorganization of the website itself. The project has since opened to other steady contributors, including a Scottish doctoral student who wrote biographies of the authors she discovered, an American postdoc who categorized novels into genres and wrote summaries, and a retired English astrophysicist who identifies our authors through Ancestry.com. We continue to invite more researchers to join us in this work.
The last several years, then, have witnessed a second growth spurt in the project. In raw numbers, the total number of titles in the database passed 20,000 titles in December 2020 and 24,000 titles in December 2023. The current website—the sixth version since the original site in 2007—is easier to navigate and looks better than ever.16 We have an ongoing plan for adding more titles and authors to the database. More important, the size of the database begins to provide real insights into the publishing history of the Victorian novel. Besides the steady annual increase in the number of titles published, the data trace trends in authorship, the rise and fall in the popularity of genres, and the growth of serialization practices, among other things. Not just big data, but big, curated data. Looking back, I am struck with how the project took on a life of its own. Short of obsession, I still find satisfaction in adding a new title to the database or identifying a hitherto unknown author. The work has become a routine part of my life, filling the gaps between teaching, attending meetings, or tending to other research projects—though, at times, it does intrude into those areas as well.
After the indulgence of a personal narrative, I will conclude with a few observations about the affordances of digital research gleaned from my experiences. First, digital research offers a scale rarely found in print projects. Even massive multi-scholar projects such as the Wellesley Index to Victorian Periodicals (1966–1989) or the third edition of the New Cambridge Bibliography of English Literature (1999) faced physical limitations as to what periodicals or authors, respectively, could be included.17 Not so for the digital, where even now storage is effectively limitless. The original database of 1,400 titles in 2007 measured less than a megabyte in memory; the current database of 24,000 titles only measures 5.8 megabytes by comparison. The finished database in the future will likely still fit on a memory stick. The only constraints on the database—or any database—will be the ones we impose: our definition of a “novel,” the historical period covered, the geographic range of our search.18 However, every virtue has a corresponding vice. In this case, the limitless breadth implies that we could be adding to it forever. Even on the day we add the “last” Victorian novel, we will continue to find typos to correct, authors to be identified, or new ways to code the data. At a minimum, the website itself will need to be maintained. Whereas print is permanent, the digital is evanescent, prone to erasure through technical obsolescence or not paying the hosting bill.19 ATCL, like all DH projects, will never be finished. Archiving and preservation of the project have an urgency that never crossed my mind when I started.
Second, digitally available research reaches a worldwide audience. As a graduate student, I often heard the quip that only a small number of readers ever see the humanities research published in narrowly defined academic journals. Even in the digital age, most academic articles and books live behind paywalls such as JSTOR and Project MUSE, effectively closing them off to the wider public. But due to the particular design of ATCL (not quite appreciated by me at first) and it being freely available, Google could log every page—including the individual title or author page for every entry—and include all of ATCL in its search database, which in our case is currently close to 40,000 pages.20 Within the first year of going live, I began to receive emails from users who stumbled across the site through googling. One of the first such emails came from an amateur genealogist who identified her ancestor as one of the authors listed in the database and asked if I would like to know more about her. (Yes, I would!) Rarely does a month pass without a descendant of a Victorian author sending me an email inquiry about their ancestor, either asking for more information from me or (better still) offering me more information. In quite a few cases, family members have identified otherwise anonymous or pseudonymous authors, sent me photos of their author, or shared treasured family copies of their books with me. Book dealers or owners of Victorian novels have been another source of emails; most want to know the value of their old book (sadly, much less than they think). Fellow scholars make up the largest source of correspondence, ranging from polite corrections to requests for bespoke datasets. I have received emails from every continent except Antarctica. This larger audience ranks as the most pleasant of unintended consequences.
Third, digital research has transformed from a domain demanding technical knowledge or expensive hardware and software into a field open to anyone. In reading over my narrative, I am struck by how little money I have spent on creating this website. The web programming book, for which I paid at most $30, taught me all I needed to know and more than paid for itself a thousand times over. On YouTube, thousands of videos will teach you programming for free. If you have learned a foreign language or mastered English grammar, then you can learn a web scripting language in which all depends on good spelling, punctuation, and syntax. The database software MySQL was free to download. The cheapest laptop contains enough computing power to manage even large databases. That said, time and labor are the biggest investment. To add one new fiction title to the database involves only a few simple steps—identifying a title in The English Catalogue, finding a copy in Google Books, typing its entry into the database—which can take anywhere from two to ten minutes. All told, adding one title to the database takes on average about five minutes. If nothing else, the data entry of the current 24,000 titles represents some 2,000 hours of work, or 250 eight-hour workdays. I may have spent a comparable amount of time programming the website and writing author biographies. So, it may be cheap, but it has been time and labor intensive.
Ultimately, digital research creates a responsibility. The analogy of parent and child may be trite, but it aptly describes my relationship to the being I created. Up to now, I have nursed it along and kept it alive, feeding it titles and authors, watching it grow, and witnessing it surprise me again and again. As its creator, I want ATCL to thrive and survive out in the wider world, a task grown easier by bringing together a family to support it. I hope that I or someone else will continue to keep the server running. To be honest, like an anxious parent, I frequently google the website to make sure it is still there. My baby.
Notes
Peter Garside, James Raven, and Rainer Schöwerling, eds., The English Novel 1770–1829: A Bibliographical Survey of Prose Fiction Published in the British Isles, 2 vols. (Oxford: Oxford University Press, 2000); Peter Garside, et al., British Fiction 1800–1829: A Database of Production, Circulation and Reception, https://www.british-fiction.cf.ac.uk; and Ebbes, Verena, et al. “The English Novel, 1830–1836: A Bibliographical Survey of Fiction Published in the British Isles,” Romantic Textualities: Literature and Print Culture, 1780–1840, November 21, 2016, https://www.romtext.org.uk/resources/english-novel-1830-36/.
Found at https://www.victorianresearch.org/atcl.
As of this date, there is no firm estimate of how many novels the Victorians published, with estimates ranging from 35,000 to 60,000 titles, so at worst about a third finished.
Fourteen years ago, the author met Garside at a conference in Edinburgh. When told of the project, Garside enigmatically quipped, “At least you are still young.”
Troy J. Bassett, “The Production of Three-Volume Novels, 1863–1897,” Papers of the Bibliographical Society of America 102, no. 1 (2008): 61–75. Readers should note that the figures in the database supersede this article.
Several people over the years have fruitfully followed this advice, including Marie Léger-St-Jean.
True as it is now, for many years I did not consider my project as “digital humanities” at all until several discussions with Katherine D. Harris finally convinced me. I thought I was just doing enumerative bibliography with a laptop.
Reading them, my then dean asked why PFW wasn’t hosting the website, to which I replied that they had had their chance.
John Sutherland, The Stanford Companion to Victorian Fiction (Stanford, CA: Stanford University Press, 1989). I happily add that John has been a great inspiration and supporter from my earliest days.
Andrew Maunder, “Bibliography of Sensation Fiction,” in Varieties of Women’s Sensation Fiction: 1855–1890, vol. 1 (London: Pickering & Chatto, 2004): 279–392.
Troy J. Bassett, The Rise and Fall of the Victorian Three-Volume Novel (Cham, Switzerland: Palgrave Macmillan, 2020).
To be fair, my nine-month academic position includes a quarter-time research release but without pay in the summers when I did much of the work.
Eventually published as Allen Riddell and Michael Betancourt, “Reassembling the English Novel, 1789–1919,” CA: Journal of Cultural Analytics 6, no. 1 (February 2021).
The latter two have appeared as Allen Riddell and Troy Bassett, “The Class of 1838: A Social History of the First Victorian Novelists,” Mémoires du livre / Studies in Book Culture 11, no. 2 (2020); and Allen Riddell and Troy Bassett, “What Library Digitization Leaves Out: Predicting the Availability of Digital Surrogates of English Novels,” Portal: Libraries and the Academy 21, no. 4 (2021): 885–900.
Marie Léger-St-Jean, Price One Penny: A Database of Cheap Literature, 1837–1869, 2021, http://www.priceonepenny.info/about.php.
Credit goes to former PFW student Vanessa Drennen for the current overall design of the website.
Eileen Curran continued to print additions and corrections to the print version of the Wellesley, work now continued by the digital project the Curran Index to Nineteenth-Century Periodicals, https://www.curranindex.org. Patrick Leary gives a wonderful history of the Wellesley in “‘I have just had one of those large ideas’: Walter Houghton, Richard Altick, and the Origins of The Wellesley Index,” Victorian Periodicals Review 54, no. 2 (2021): 279–303, which coincidently has many parallels with my narrative.
I am already reining in Allen, who wants to expand into America and the European continent.
Our website crashed five years ago and became inoperable when the hosting company updated their server software. It took a concerned user to alert us, and we spent several weeks updating our code in order to get back online.
Recently, Wikipedia discovered ATCL and began linking their entries to ours.
