Let's start with a truism: not every book is suited to electronic publication. We can arrange titles along a rough continuum. At one end are the most bookish of books: Tristram Shandy, for instance, is so obviously a product of the printing press that it is irreproducible in any other medium. At the other end are the most referency of reference books, whose sole value is their information, irrespective of the form in which it's delivered. There's no reason whatever, to take an extreme case, to put the three billion pairs of A's and C's, G's and T's of the newly decoded human genome in print: readers, if we can call them readers, want to make rapid sorties, get what they need, and get out of there. Computers are perfect for that sort of thing.
The question is where CBEL falls on the spectrum. I'm not convinced it's a "pure" reference book, like a list of DNA base pairs; there's much to be said for the pleasures of browsing and serendipity, both products of the print world. Still, whatever the future of digital novels and Rocket eBooks, the computerized reference book is here to stay. Some paper reference works are already obsolescent: consider, for instance, the union catalogue. The current generation of graduate students will probably be the first never to handle the National Union Catalogue's 800 volumes, or the massive printed catalogues of the New York Public Library or the British Library. Though abandoning such works altogether is premature much has not been digitized it's just a matter of time before the books are wholly superseded; and though the nostalgic part of me will be sorry to see them go, other parts say good riddance. The disappearance of printed reference books may please us, depress us, or leave us indifferent, but it's probably a fact with which we have to reckon.
I suspect, then, there's a degree of inevitability: soon everything is likely to be available electronically, either exclusively or as a companion to a print volume. So rather than debating whether, let's turn our attention to why. It's easy to catalogue advantages of electronic reference works, but I'll start with a big one: electronic publication makes possible much more frequent updates than books did. The CBEL series is averaging two or three editions a century, a pace obviously determined by the world of print. With some future E-CBEL, there's no reason it can't be two or three editions a year, or even two or three editions a month. It all depends on what we mean by "edition": the whole notion of editions with ordinal numbers may be a relic of print culture that vanishes in an electronic age. OED3, for instance, may well be the last one to bear a number; later we may see only OED-of-the-afternoon-of-the-sixteenth-of-June-around-four-o'clock. The benefits of such currency are obvious, though whether Shef Rogers, his team of contributors, and Cambridge want the headaches of managing such perpetual flux, a Maoist permanent revolution, is an open question.
Notice how quickly advantage shades into disadvantage: greater flexibility of product demands greater flexibility of producer. And while I'm listing problems, I'll mention another: money. On-line publication offers some savings, particularly in materials; a CD costs a buck to press, compared to the considerable expense of paper and bindings for multi-volume reference works. But it also introduces new expenses, and besides, the system by which people buy books is old and familiar. What about less tangible works? Site licenses are the means publishers have worked out to collect money for Internet-based resources, though pricing schemes are still up for grabs. The danger is that those without institutional support will be shut out of the digital future: those without university IDs, maybe even those with IDs from less affluent universities, may be left in the cold. Granted, the CBEL is consulted almost exclusively by academics, but there are plenty of unaffiliated scholars. In the old days, they could use expensive books in a library for free. But resources linked to student and faculty ID numbers may make research impossible for them. Some publishers have experimented with individual subscriptions and pay-per-view plans, but prices are almost always beyond the means of private researchers. The stratification of the academic haves and have-nots is already too pronounced, and I worry that the brave new world will only make it worse.
Having raised a few possible advantages and disadvantages to moving a familiar book to an unfamiliar medium, I'd like to discuss how the unfamiliar medium might influence the familiar book even before it becomes an electronic book. Every revolution in technologies of the word has brought about changes in the organization of information. Reference books are products, perhaps even causes, of literacy itself: in most cultures, the earliest written matter is simply lists, like the large majority of tablets in Linear B or cuneiform. New advances like alphabetical order, page numbers, tables of contents, and indexes made it possible to organize old information for new purposes and, as a side effect, revealed the limitations of the old technologies. As we begin thinking about new ways we might use an old book, we discover new things we'd like to do with it, new ways of searching it. But these things require plenty of forethought.
Some examples. The electronic transcription of OED2 is a godsend, but it doesn't yet take full advantage of the electronic medium. The late Victorians did not design it with electronic searching in mind, and things like inconsistent abbreviations make some dependable results impossible: you cannot reliably find, say, every Shakespeare citation, because Shakespeare is sometimes cited by name, sometimes by abbreviated name, sometimes by title, sometimes by abbreviated title, and so on. Johnson's Dictionary, now available in a magnificent Cambridge CD-ROM, also provokes frustration: a search for legal terms, for instance, fails to turn up law, because Johnson couldn't imagine a reader too stupid to recognize law as a legal term.
Computers, though, are that stupid, and this forces the designer of a new reference work to try to imagine the ways someone might want to retrieve information. The search mechanism is easy enough to change over time, but only if the underlying database is designed with these things in mind. Someone may want to see all the political pamphlets published by women in Edinburgh in the 1780s, or all the poems that went into third editions within ten years of their original publication. If that information isn't consistently encoded in the database from the beginning, the query will be impossible, the question will go unanswered, and a valuable opportunity will be wasted. There's no guessing what topics future scholars may find interesting perhaps many researchers in 2051 will want to search for books by left-handed authors or books bound in red morocco, and will curse us in 2001 for not having the foresight to put such information into our reference books. Perfection is not to be expected in the sublunary world. But we do have an obligation to make our best guesses about what might profitably go into a bibliography of English literature.
Shef tells me the plans are underway to prepare the underlying database for this CBEL volume in XML, a markup language in which various kinds of information are tagged, and which then allows searches based on the structure of the tagging system. That means that now is the time to think these things through, not only for CBEL's possible digital future, but for any reference work. It's a chore to design a coding system from scratch, but it's a thousand times harder to go back and retrofit a poorly conceived work.
I suppose what I'm asking is how we should conceive of a book that's also a database, even if the database won't be publicly available for some time. Books never had to be databases before, but we're in the uncomfortable position of trying to create a hybrid that won't foreclose any possibilities that arise down the line. It's impossible that all these questions will be settled to anyone's satisfaction before this volume of CBEL3 goes to press in another five years or so. But now's the time to begin thinking about them, and I hope that in the discussion to follow we can bear in mind that the new technologies offer new freedoms, as we're allowed to reconceive how we should use our familiar tools but they also impose on us new responsibilities, as we're now forced to reconceive how we should use our familiar tools. I'm pleased, therefore, that Shef Rogers has invited comments on what his volume should look like, because it's much better to worry about it now than to complain about what could have been.