1988-1992  •
1993-1996  •
1997-2000  •
2000-2005  •



History of the Project

One thing leads to another. Such is the unmethodical way in which traditions and archives evolve: one thing leads to another, and another, and before long things begin to assume aspects quite unforeseen at the outset.

I relate the history of this project less as a cautionary tale (no one would now make the missteps that I took) than to illustrate how such a process works, and to explain aspects of ENGLISH POETRY 1579-1830 that might otherwise seem odd. Things would have been done differently had I known at the outset what was learned along the way. Since the begetting of this archive was contemporaneous with the introduction of digital texts in the humanities, a retrospective view may be of some interest as a record of how one archivist responded to the sweeping changes affecting the field.

There are really two narratives here, a story of what was learned about the circulation of information in and about literary history, and a story of what was learned about the organization and distribution of information in digital texts and databases. Since each has had an impact on the other what follows might be styled "The Confessions of a Digital Antiquary."

1988-1992. Beginnings.

Research began very soon after I arrived at Virginia Tech in 1987 following my graduate work at the University of Virginia. While converting my dissertation into a tenure-book on georgic poetry and meditational writing, I was led to reflect on the origins of modern ideas about culture that seemed profoundly at odds with conceptions of order and change in georgic poetry — an obvious place to look for the origins of cultural discourse. When and how did cultural discourse evolve? I began a study of eighteenth-century imitations of Spenser, thinking that there I would find the nexus of ideas about history, nationality, imagination, and education that gave rise to nineteenth-century ideas of culture and modern conceptions of literary history.

In these imitations I found one of the paths to modern historicism, but also a new line of inquiry. Cultural discourse seemed to originate as a kind of fairy fiction, a way of writing about origins and tradition that treated complex historical processes as imaginative wholes. While talk about culture emphasized originality, the poems themselves were obviously products of imitation. The "culture wars" being then in full force, I was inclined to be skeptical about claims being made for and about culture. Georgic conceptions of historical identity seemed more persuasive: they relied less on sweeping metaphors and holistic paradigms more on discrimination among concrete historical instances. I set out to make a complete enumeration of Spenser imitations.

Since there were quite a few imitations, most too obscure to be treated in any detail by literary historians, and since it was apparent that there was little consensus about what constituted an "imitation," there was much bibliographical work to be done before the necessary discriminations could be drawn. I spent a couple of years going over the Spenser literature, spending time at the Clark Library in a seminar with Robert Folkenflick, and at the Houghton Library in a seminar with Leopold Damrosch.

The original bibliography took the form of note-cards wrapped with a rubber band: there was one card per work, listing title, author, date, citations in the literature, and a brief quotation to help me remember which work was what. One could arrange the cards in a continuous series, or pull out a subset to define series within the series. I was familiar with word-processors, having in the course of writing my dissertation made the move from an ancient IBM typewriter to a newfangled Wang word processor. Yet pencilled note-cards remained the more efficient technology for cataloging, at least until the back of a card became filled up. Adding cards for a single record, albeit carefully numbered, was not a very efficient way organizing information.

By 1990 the note-cards had been transferred to a WordPerfect document on a Tandy laptop running MS-DOS. Here was nigh unlimited space! I began learning about cataloguing the hard way, using a non-arbitrary numbering system (1749.1, 1749.2, 1749.3); this caused no end of ills; when my sources got dates wrong or multiplied entries for single titles (as often happened) adding or deleting records involved renumbering the series. The notion of using a database for this work was as yet foreign, at least in English departments, at least to persons like me. At the time it seemed just thrilling to do keyword searches: with only a few seconds of chugging, WordPerfect could scroll through hundreds of pages of text to find just the item requested.

What was yet more exciting, one could make a print-out from the digital document and so admire the material results of one's labor. Before long, I had a 300-page annotated bibliography, attractively formatted and reposing handsomely in a three-ring binder. I made frequent print outs (200 pages, 250 pages, 300 pages!) But why consult the printouts? It was easier to use the digital document. Nonetheless, with the Web as yet several years away, scholarship was still all about paper. The tenure process was certainly about accumulating paper: by Fall of 1992 I was able to muster the canonical book, articles, and reviews, looking big enough when photocopied into an impressive heap of pages. The three-ring binder, however, was not submitted.

1993-1996. Pastoral Interlude.

Tenure brought relief from the ticking clock and the illusion of endless vistas of time. Why not undertake a database of Spenser imitations as a research project? I was accustomed to financing my own research, and it was possible to purchase the necessary software and a primitive though costly scanner. My introduction to databases came when I migrated my text file to an early version of Filemaker Pro running on a Mac Quadra. The graphic interface was a novelty, and much time was spent modifying the visual arrangement of the fields and the color of the fonts. These new technologies — scanner and database — changed the scope of the research. Rather than transcribing a couplet or stanza as an "illustration" it became easy to input entire documents. Given the antiquary's predilection to accumulate, it was all but inevitable that the bibliography would evolve into a full-text database.

This was not planned; it just gradually happened. I also added a second file with records for authors; there would have been four or five hundred then, and it was becoming difficult to distinguish the Rev. John Whaley (d. 1745) from the Rev. John Whalley (d. 1748). Without knowing what I was about, I had made the first step towards a relational database. But I did not know what a relational database was, and moreover I found the Filemaker scripting language unintelligible. A computer-savvy colleague told me about a Mac program called HyperCard that would make scripting intelligible and easy. This proved to be true, and having spent only a few months in Filemaker I began the work of cutting and pasting the data into HyperCard "stacks." The note-card metaphor implicit in the HyperCard software made the transition to database design remarkably easy. Too easy: it required little conscious thought, and I should have been thinking very carefully about designing a relational database.

But here was a relational database of sorts: each "stack" could share data with the others, a wonderful concept suggesting possibilities far beyond paper, or text files, or even a simple database. I was learning to make texts interactive. Not knowing the first thing about databases, I made a fundamental error in design: I began collecting "illustrations" for author records as I had earlier done for text records, pasting them into fields. They should have had their own records of course, but I avoided this by devising a clever script to display the illustrations without having to scroll through their fields. Clicking on a header excerpted the desired chunk from its hidden field and displayed it on the "card."

This was a bad, bad thing to do, an error in database design that would later cause much gnashing of teeth. But considered in its own right it was wonderful. The script I wrote was like a clever little sonnet, remarkably tidy and cleverly self-recursive. It was my introduction to the mysteries of text-chunking, link design, and scripting. Consider that the World Wide Web was still a year or two away, and imagine my excitement as I first grasped the possibilities of using linking digital documents to map a literary tradition. By 1992-93 there was already much talk about linking documents, and several products designed for this purpose were being tested in the English department.

Having digested most of the Spenser scholarship into the database, I was so fortunate as to have a book contract. Writing Edmund Spenser: a Reception History provided the necessary impetus to shift from accumulation mode to interpretation mode. Spenser was a happy object for such a study because the critical record for this poet is nigh coterminous with the history of criticism in English. One cannot but be struck, or at least I couldn't, by how radically modes of criticism had shifted over the centuries, out of verse and into prose, then out of the magazines and into academic books and journals. Moreover, so it seemed, Spenser had an impact not only on what critics said, but how they said it. There were subtle connections between forms of poetry and forms of criticism that needed consideration.

Commentary about Spenser evolved as later writers selected new elements of his poetry to comment on, but also as they invented new ways to incorporate those elements into their modes of writing: seventeenth-century epigrams, eighteenth-century allegorical odes, nineteenth-century biographies, all imitated Spenser as well as commenting on him. Something like these historical differences was at work in the academic philology I was quarrying for information: scholars of Elizabethan, Augustan, and romantic poetry selected different elements for comment and wrote about them in different modes for criticism.

To scholars of renaissance literature, bless them, every blot was important and necessary to be recorded and described. Augustan scholars were more polite and more casual: if one or two members of the literary club went missing there was no great loss provided the group was present. These critics could be very condescending towards minor writers, though nothing compared to scholars of romantic literature who sometimes proceeded as though minor writers never existed (I am speaking of an earlier age of scholarship). Elizabethan poems were few and all precious; nineteenth-century poems were thick on the ground and discrimination was deemed essential. This resulted in a curious state of affairs: while all agreed that Spenser was terribly important in the romantic era, bibliographic research on Spenserianism all but terminated at the year 1800.

There was more. Critics of romantic poetry tended to assimilate romantic ideas about culture, among them very undiscriminating ways of thinking about relations of parts to wholes. Where organic connections can be assumed, complete enumerations are deemed unnecessary: find the best or most paradigmatic part, the essential moment, and you have it all. As a skeptic about cultural discourse it was my business to count heads and draw distinctions, if only to keep my data consistent across the two and a half centuries covered in the archive. Since the publication of the Spenser book in 1996 much time has been spent patiently (sometimes impatiently) working through the back-alleys of romantic literature so that traditions developed out of Spenser could be made visible in the whole rather than as a whole.

As the archive took shape, the Faerie Queene suggested alternative ways of thinking about order, as it always seems to do. Like a database, Spenser's poem was encyclopedic, grew by accretion, and was never destined for completion. Its design was subject to revision in the course of composition. As the several books of the Faerie Queene display marked modal differences (being alternately and inconsistently pastoral, satirical, burlesque, heroic), so do the several eras of literary history documented in the archive. The series of poems comprising its traditions behave like the threaded tales in Spenser's romance: pastorals, sonnets, allegories take the center stage, recede, then turn up in a different book partnered with others. They are neither autonomous nor integrated; possessing degrees of independence, they remain parts of a larger design, of several larger designs. It was no great stretch to think of the archive as a digital land of faery recording the acts and utterances of literature's renowmed knights, subtle doctors, brilliant ladies, and irrepressible clowns.

To render the archive's maze of linked paths and nodes interactive, recourse to a scripting language was necessary. HyperCard used a scripting language called "HyperTalk" cleverly written to resemble English. While the use of prepositions and articles might seem a little odd, there were at least recognizable nouns and verbs. Write in the command line "Go to next record" or "Show field bibliography" and the genii worked its magic. HyperTalk could boast a suite of text-handling functions now divided among SQL, PHP, JavaScript and CSS, with others besides. The years 1993-1996 were spent exploring this new language: making an "archaism-finder," click-and-show annotations, elaborate search-and-display schemes. Like Adam's language in Paradise, HyperTalk seemed powerful, effortless, and comprehensive.

While HyperTalk's text-chunking capabilities were splendid (one could refer directly to "characters," "words," and "lines"), the software was in other respects primitive. Designed for the first generation of Macs, the tiny "card" image displayed was scalable only with difficulty; color did not work particularly well. The visuals native to the application were nothing if not ugly. In earlier versions of HyperCard formatting was lost when shifting text from one place to another, so I simply dropped formatting from my data-entry. There were also more serious, if less visible, problems with HyperCard, though in my innocence I failed to notice, or at least to comprehend, its fatal limitations.

Chief among these was the fact that HyperCard was not cross-platform. This did not seem like such an issue at a time when applications were seldom cross-platform and the Mac was riding high. But while I was happily scripting digital dungeons and dragons the World Wide Web came online and cross-platform computing became a priority. I made my first web-pages in 1994, but was not much impressed with HTML. While more thoughtful persons recognized immediately that this was the future of digital texts, static web pages held no enchantment for me. Also in 1994 I attended a fine seminar on making digital texts run by David Seaman at the Rare Books School in Charlottesville where I was exposed to SGML and the proper way of doing things. But in 1994 the possibilities for making and managing web pages were very limited: the idea of managing a collection of thousands of hyper-linked web pages had very little appeal.

So I persisted in my youthful errors, expanding the project by adding elaborate glossaries, bibliographies, various kinds of indexes. Click the word "whilom" and poof! there would appear a gloss for the term with a list of all the Spenser imitations that used it. I could count lines; I could indent stanza patterns on the fly. What HTML document could do such things? Rumor had it that a Web-compatible version of HyperCard was in development, and I had scarce a worry in the world. In 1996 I gave the first public demonstration of the project to the English Department under the auspices of my good friends at the Center for Alternative Technologies in the Humanities (CATH) at Virginia Tech. It ran on a set of networked Macintosh computers.

1997-2000. Georgic Labors.

I should have been very worried. But my concerns then had less to do with distribution than with editorial matters. The project had been conceived as an annotated bibliography with added "illustrations." It was now in the process of becoming a full-text database. With little knowledge of the the editor's craft, I was at first reluctant to commit myself even to transcription. The problems were formidable: there were issues of copyright (especially unclear in the early days of the Web). There were issues of using original or modern-spelling texts. There were issues of text-formatting and the use of special characters that would be portable across platforms. Prudence suggested that textual editing be left to those who knew the business.

But I was not prudent. Digital technology made text entry remarkably easy, and since my project was more concerned with relationships among documents that with representations of documents, I elected to go with plain-text transcription and minimal editorial intervention. When the Chadwyck-Healey English Poetry Database came online in 1995 I envisioned my archive as a supplement to, not a substitute for, that sort of HTML-based digital archive. As a practical matter, I needed to find a compromise between speed and accuracy in making transcriptions. My early efforts were appalling: there were missing lines and garbled words; first efforts at OCR produced hilarious errors. I soon learned that I could type and scan only so many hours a day without suffering pain in hands and eyes.

The key to working any large project is learning to manage time: alternating typing and reading permits longer sessions, alternating time spent researching and transcribing keeps the mind fresh and alert. One learns that even the longest and trickiest texts are manageable given time and patience. One learns to break the labor down into small parcels that can be handled amid the other routines of life. There are seasons for all things: the everyday work of data entry was punctuated by occasional sessions digesting reference materials, building scripts, compiling indexes, and formatting the database.

Large projects are usually undertaken by consortia of specialists who divide the labor according to plan and method. While this creates obvious efficiencies, there are advantages to going it alone. As the project designer learns more about the subject in the process of data collection and transcription, the database can be modified accordingly. While it can be difficult and possibly disastrous for a group to make large changes in midcourse, a solitary antiquary working to his own schedule is able to begin again and to do things for a second or third time without violating a schedule or holding up the work done by others. If it all takes longer, the laborer knows that wisdom is the daughter of experience, and experience the daughter of time.

One thing I learned by experience was the importance of periodical literature to English poetry. I had become more familiar with magazines in the process of seeking reviews that might be culled for headnotes or lead to neglected poems. Magazines were also fertile sources for Spenserian poems overlooked by previous bibliographers. In gathering and transcribing periodical verse it became apparent magazine ephemera had a vital role to play in shaping and reshaping literary practices and traditions. Moreover, the very concept of a periodical "magazine" suggested new ways of thinking about the database.

Many periodical writers were already familiar as authors of minor Spenserian poems. In fact, eighteenth and nineteenth century imitators of Spenser and Milton seemed to predominate among reviewers, editors, anthologists, and memoirists. From these same magazines could be gleaned more information about these minor figures than one might suppose. Rather than quarry the periodicals for facts in the manner of a reference work, it seemed best to collect the documents themselves. It is one thing to know that poet A knew poet B, but a document recording a supper invitation or memorializing a friendship opens a view on the world. Much of the information I was discovering about personal and professional relationships was recorded in uncollected periodical verse. I began entering sonnets and verse epistles, obituaries and memoirs of living authors. As periodical material accumulated in the commentary and biography fields, series and sequences began to appear not unlike those being recorded in the text file; "tradition" was as busy with the poets as it was with the poems.

With the introduction of the Web, the word "information" began to figure ever more largely in conversations about literature. As a confirmed antiquary, a diligent pursuer and hoarder-up of forgotten facts, information had long been my bread and butter. Laboring on the database reinforced the antiquarian proclivity to hoard and accumulate: empty fields were an abomination; a date of birth must be recorded, publications found, criticism located, a list of acquaintances compiled. Magazines were the mother-load of obscure information. The word originally meant "storehouse," and parallels with the new media used for managing information were not difficult to draw: the Gentleman's Magazine was like an eighteenth-century version of the Web, a place where correspondents met to exchange information on every conceivable topic. It was also rather like a database, slotting information into a regular, open-ended structure of records, fields, and indexes.

It says something about the relationship of poetry to information in the early-modern era that poems appeared in periodicals mixed in with reviews, obituaries, articles on current events and literary history. The same names and topics would recur in all, often juxtaposed on the same page. The information I was collecting had once been consumed by the very poets whose lives and writings I was recording. Their poetry, much of it topical and uncollected, was often replete with information about persons, places, and things. Readers of those journals, many of them poets, and many of those poets editors, anthologists, and memoirists, were plainly obsessed with collecting information, and in particular with information connecting the literary to the educational, social and political spheres.

Adding this ephemera to the database made the contents more magazine-like, and also underscored the relation of information to time. Information published in periodicals (as opposed to dictionaries and encyclopedias) comes with a date-stamp. It is expected to be new, or at least novel. Chronological sequences of documents extracted from periodicals — poems, anecdotes, and memoirs — registered the progress of taste and mapped the progress (or decline) of knowledge about specific writers and their poetry. Material added to the commentary and biography fields made the project a more useful tool for assessing the evolving state of literature: who was up and who down, but also (when given in the form of documents) how literature shifted its priorities and modes of communication as time passed.

The daily and weekly work routines involved with assembling the archive resembled the labors of periodical editor in quest of fresh matter. The ideal was a hour reading microfilm in the afternoon followed by two hours of transcription in the evening, with larger sessions reserved for weekends. The regular routine would be punctuated by occasional packets from interlibrary loan. I would concentrate work on a particular chronological era for months at a time: a new poem might lead to working up a new writer, and connections to other contemporary poems and writers. When boredom threatened, I could shift to a different epoch or geographical area and work there for a while.

So regular did these habits become that it is now difficult to remember what was done when: one year was very like another. The routine was occasionally punctuated by more memorable events, like doing a presentation in 1997 at the Digital Resources in Humanities Conference at St. Anne's College in Oxford. That was a high point: I received some flattering attention and even had a conversation with a commercial vendor. But in 1999 I hit bottom at the ACH-ALLC International Humanities Computing Conference in Charlottesville, where no one seemed remotely interested in the clever things my database could accomplish. This was partly because the ACH-ALLC is deeply committed to SGML, but chiefly because by 1999 HyperCard had become an outmoded technology. Digital texts had moved decisively to the Web.

Rumors of HyperCard's impending demise had been circulating for a year or more. The proposed upgrade, HyperCard 3.0, seems to have been terminated by Apple in 1998, and while this was denied, a "Save HyperCard" effort was producing no results. By 2000 my software application was in trouble and something had to be done.

2000-2005: Epic Proportions.

There was nothing for it but to begin again. It was possible to save the data from the shipwreck of the database of course. The last HyperCard script I wrote walked through the records, writing them to tab-separated text files. As the script ran and the counter turned over I thought of an orderly march of refugees abandoning a doomed city. Troy was in flames and Troynovaunt nowhere in sight.

My colleagues at the Center for Applied Technologies in the Humanities were using a Filemaker database for the Gravell Watermark Project and suggested that I follow suit. I was now prepared to be more patient than I had been before, and after all, Filemaker, like HyperCard, was designed for use by non-programmers. Unlike HyperCard, Filemaker was being retooled to work on the Web. The new software made database architecture easier to comprehend and became the occasions for a major improvement: rather than storing biographies and commentary in concealed fields, each biography and comment would have its own record, linked to the author record by a key. Creating these new records involved forty hours of by-hand cutting and pasting data, labor that would not have been necessary had I done things properly to start with.

The new scripting language was, compared to HyperTalk, a very blunt instrument. While database operations were easily managed, text-chunking required multiple nested functions in code that looked barbarous and was torturous to write. After some months of working with the scripting language I was able to recreate but little of the former functionality: archaism-finders and annotation-makers were out of the question; no longer could users click on "pastoral" and magically travel to the next instance of the kind. While this was disappointing, simplicity also brought clarity. I concentrated my energies on adding content rather than functionality.

The archive was broadening from Spenser imitations to the larger network of poems, criticism, and biography in which they were imbedded. It had long been apparent that most "Spenser imitations" were really imitations of Shenstone's Schoolmistress, Thomson's Castle of Indolence, Beattie's The Minstrel, Burns's "Cotter's Saturday Night," and so forth: series within the larger series of Spenserian poems. It was becoming apparent that series of imitations evolved in ways peculiar to themselves, ways related to the topics under discussion in a particular series. It was desirable to have more groupings available for analysis and comparison. I also noticed that there were more poems in variants of the Spenserian stanza than in the regular form, and that the variants were often used for Spenser imitations; until the variants were documented the status of the Spenserian stanza would remain unclear. It seemed that scholars more familiar with nineteenth than with seventeenth and eighteenth-century poetry had overestimated the iconic value of the nine-line stanza. On the whole, it seemed that use of the pastoral eclogue was a more reliable marker of Spenserianism than use of the stanza.

But was establishing a more rigorous definition of Spenserianism really the point? By the 1740s when the imitation of English poets was getting seriously underway, Spenserianism and Miltonism had become thoroughly entwined; moreover, the emergent series of imitations had less to do with Spenser or Milton per se than with attempts to make English poetry "British" (or in some contexts, Scottish, Irish, or American). If this was not already apparent from the style, it was certainly apparent from the subject matter, especially in the uncollected topical verse where Spenser and Milton were frequent objects of imitation. The oldest and most "traditional" manner was being adopted for the newest and most ephemeral subjects, a seeming paradox underlying the fairy-fiction dimension of emergent cultural discourses that concealed their essential contemporaneity under a cloak of antiquity. It was time to make the leap from magazine verse to newspaper verse, the most ephemeral poetry of all.

Not that all newspaper verse was ephemeral. Many poems bibliographers had sourced to collected works, magazines, and anthologies actually made their first appearance in newspapers, where they could be found among other, uncollected examples of Spenserian and Miltonic poetry. If newspaper verse was the bottom of the literary food-chain, it was also in some ways foundational to the larger economy. Newspapers, which occasionally printed poems in Latin, French, and even Greek, were obviously consumed by members of the literary establishment and served as conduits for information about the state of the literary system. For purposes of comprehending traditions in English poetry as process and not just as product, the diurnal prints could be as important the landmark anthologies. It was through newspapers that much information about the literary system first got into circulation, and through newspapers that much of the business side of literature (ignored or reviled in cultural discourse) was transacted.

As I was starting down this path in 2001-2002 the news media was filled with comments (generally hostile) about the "24-hour news cycle." The Web was supplying news and views that challenged the authority of traditional outlets in metropolitan centers. Had something similar transpired in eighteenth-century poetry when rootless periodicals displaced the court and universities as centers of literary production? Periodical poets were also cutting into territory previously occupied by broadside ballads, the traditional source for topical verse. From the 1740s forward, readers could follow the news in diurnal and hebdomadal verse, panegyric or burlesque as the case might be, but in literary as opposed to demotic forms. Military victories, political scandals, celebrity deaths, and court affairs provoked torrents of periodical poems from all corners of the English-speaking world. Nor was literature ignored: newspaper "bloggers" commented on poets and poetry with a candor, enthusiasm, and specificity often lacking in more official reviews.

The inclusion of periodical ephemera swelled the size of the archive in all its dimensions. In the years from 2000 to 2005 the number of records more than tripled as gaps were filled in and a serious excursion was made into American poetry. By 2004 the number of commentary records had surpassed the number of text records; the project was now as much about reading as about writing. As the commentary records swelled, it became easier to get a quantitative as well as a qualitative grasp of who was being read, when, where, and how. As more poems and series of poems were added to the archive the distribution of records over time began to shape themselves into meaningful statistical curves.

All of this was very exciting. Yet until the hoard of information could be released onto the Internet the archive would have no readers. Having made the migration to a proper database, the next step was to master the business of creating dynamic pages to port it to the Web. As anyone who has gone through the process knows, making pages dynamic involves a conceptual leap: instead of formating a document, one writes rules for formating documents. Since this requires working with multiple scripting languages and factoring in browser behavior, there is much involved.

Filemaker promised the capability of porting one's data to the Web with the click of a few buttons. This it could do, provided that one used standard templates and forewent all but the most basic search functions. Beyond that, one could use its proprietary tags (CDML) to design custom form pages. As usual, venrturing beyond the common way of doing things required experimentation. I knew better than to expect much magic from the software's "wizard" feature. Weeks were spent trying to make the HTML pages functional and attractive, and months learning to write the scripts necessary to make my already circumscribed search and text-handling functions operable on the Web.

By 2003 the project was operational. Then came a major disappointment: while all worked smoothly on my home computer, the project would hang when run using the Filemaker server software at CATH. We never could identify the source of the problem; the Gravell Watermarks database worked fine. Perhaps the poetry database was too large or the scripts too complex. More than once I experienced the classic moment of standing before an audience to demonstrate the project only to discover that there was no project to demonstrate. Here was shame; while not working with borrowed money, I was working on borrowed time. Like some Grub-street hack, I had lived for years on the promise of a great work for which there was nothing to show. I began receiving unwelcome visits from the Giant of Doubting Castle.

It was time to begin again. I gave XML a hard look, but concluded that for a project more concerned with linking documents than analyzing them the better solution was the combination of PHP (scripting language) and MYSQL (database software). Having been warned that this was not software for beginners, I availed myself of help by offering the database as a project for an undergraduate computer science class taught by Ed Fox. My five collaborators were to recreate the Filemaker project using PHP and MYSQL. They got underway in January of 2004, and presented the result of their labors the following May. There were fatal flaws, but some of the work was really very fine. Still, the term was over and the project was in an unusable state. Over the summer I found a dog-eared copy of a UNIX manual propped against my door inscribed "for your journey into php + mysql. E-mail any questions. Patrick." It had been left by one of the CS students, who with a touching lack of condescension encouraged the hapless English professor to press on.

The learning code was indeed steep. To install the software it was necessary to use UNIX from a command-line interface. While I had used a command line in HyperCard, venturing into the bowels of a modern computer was altogether more challenging. UNIX commands generated columns of unintelligible gibberish representing the secret caverns and hidden passageways through which it was necessary to pass. The absence of a graphic interface produced a sense of blindness; errant commands sent misdirected files down unfathomable black holes; security warnings impeded progress. But Patrick had equipped me with a golden bough and after two days of hair-pulling the appropriate sequence of commands was discovered: MYSQL sent its "hello" message and I was underway.

Six weeks of uninterrupted 12-hour days were required to get the site up and running: form pages, results pages, record pages. Then the challenging part began; scripting links to move readers from page to page proved to be the most difficult thing I have ever attempted. Much of this difficulty had to do with passing text through the links, text that had to be formatted and reformatted to work with PHP, SQL, HTML, and the URLs. Passing unpredictable chunks of text and symbol through a URL is no simple matter, but since PHP was designed with this in mind the proper combination of functions could eventually be found. The process resembled some of Guyon's more perilous passages: the proffered assistance of this or that function would as likely as not produce deadly errors that might lurk in concealment for months until tripped by an unanticipated comma or apostrophe in an innocent-looking piece of text.

The complexity of the scripting was such that it would take two years to work out the bugs; dozens of scripts and subroutines are required to generate any given page, and modifying any one page might produce unanticipated results in another. Mastering the coding process would be work of time, and in that respect not unlike poetry: while each script is unique, there are genres for scripting as there are genres for poetry. Code is recycled and modified in ways not unlike forms and diction in verse: from half a dozen ur-scripts most of the others have been generated by processes of adaptation not unlike the series of poems documented in the archive. But all of this is meant be invisible to users; however various the operations they trigger, the appearance of links on the page should be kept as regular and simple as possible.

Since this last migration the archive has reverted to the works-and-days routine of the former georgic regime. I have striven to refine the visual presentation of information by keeping clutter to a minimum and keeping to a clear and consistent hierarchy of forms; it seems that that this aspect, like the gathering of data, is a never-ending process. It has been necessary to find a workable compromise between complexity and ease of use in making the search pages: there are many other possible ways to parse the data, but too many choices might lead to visual confusion or worse. The work of scanning microfilm proceeds apace, and I hope soon to gain access to the full-text databases that are already transforming the ways we think about literary history.

But the potential of full-text databases will not be realized until they are properly indexed and linked to other sources of information. Perhaps the present project will suggest some of what might be done. In the mean time, I anticipate more assistance from readers who know more about the areas and authors covered in this archive than I do. Large portions of the documentary record remain to be examined, and additions will certainly be made. Perhaps in the future ways will be found to link this archive to others on the Web, making it a much more valuable resource than it is at present. One thing leads to another, and I hope that the belated appearance of this project will be no exception


One of the sadder things I have ever read is Samuel Egerton Brydges's autobiography. Angered by debt, exile, thwarted ambition and literary failure, he bitterly regrets the decades he had devoted to antiquarian pursuits. Joseph Ritson apparently died mad as a hatter, spending his last hours stuffing his trove of manuscripts into the grate and setting them aflame. Antiquarian scholarship tends to be a lonely business, plagued by its characteristic moral failings: acquisitiveness, arrogance, and melancholia. But the antiquary's quest for information can lead to happier outcomes when friendships formed in the course of the work temper the pathologies incident to the profession. I have been fortunate in my friendships.

Those who know them will recognize the debt this project owes to my teachers, Ralph Cohen and Alastair Fowler: to their work on Spenser and Thomson, but more particularly to their commitment to comprehensive knowledge and probing thoughts about time and historicity. In its early stages this project was given a boost by Robert Folkenflick at UCLA and Leo Damrosch at Harvard, and more recently by Christopher Fox during an NEH seminar at Notre Dame. Kevin Cope and friends at SCSECS have accorded antiquarianism unwonted honor, while Richard Sher and friends at ECSSS have set high standards for the sociable exchange of information. I owe a special debt to Roger Robinson of London, whose conversations about Beattie and Spenserianism I miss more than I can say. Seabrook Wilkinson of Charleston has fired my enthusiasm for local knowledge and antiquarian pursuits. My dear friend Michael Coyle has ever lent a sympathetic ear.

Closer to home, I owe a particular debt to Peter Graham, who has had to listen to rather a lot over the years. My colleagues at CATH, Len Hatfield, Dan Mosser, and Ernest Sullivan have given good advice, not always heeded. We have heard the chimes at midnight. What Spenserian would not benefit enormously from having Esther Ritchie as an office-mate? The English Department at Virginia Tech has been "open to new ideas," as they say, willing to give me the long leash required to undertake a seemingly interminable project. My students, and especially Amanda Jones and Stella Osborne, have been an inspiration. Patrick Lievsay saved my bacon at a critical moment. Eddie Watson at the Faculty Development Institute may perhaps remember some dreary hours devoted to this project many years ago.

I am grateful to the staff at the William Andrews Clark Library, the Houghton Library, the Folger Library, the Hesburg Library at Notre Dame, the British Library, and the Advocates' Library in Edinburgh for their assistance in leading me to books. The interlibrary loan office Newman Library at Virginia Tech has been extremely generous with time and treasure; the staff has been unfailingly patient with the "old books guy" who insisted on having this or that particular edition; "Mr. Bow Tie" is also grateful to the microfilm staff for keeping his favorite machine operational through hell or high water (mostly high water). Many of those who have helped me along I have never met: books used for this project have come from every corner of the United States, though I would single out UNC Chapel Hill for its excellent holdings in nineteenth-century Scottish poetry, and the University of California Davis, whose splendid collection of romantic poetry on microfilm has been an invaluable resource. To my alma mater at Virginia I am indebted for the loan of two scarce but essential microfilm collections: "The Eighteenth Century" and the later issues of "Early English Newspapers." The ESTC, on disc, and then on-line, has been an invaluable resource.

I am grateful to John Dussinger for assistance with Thomas Edwards, and to David Vander Meulen for help with the design of the front matter. Gillian Hughes has helped with correcting the James Hogg entries, and David Latané those for William Maginn.

To my wife Margaret I am indebted for my first introduction to computers and for keeping me tethered to the living world with a firm but gentle hand.

This project is dedicated to the memory of George Burke Johnston of Blacksburg, Virginia (1907-1995): poet, scholar, printer, antiquary, and gentleman.
February, 2006.

David H. Radcliffe
Department of English
Virginia Tech
Blacksburg, Virginia, USA