In 1988, when I bought my first real computer, I was much concerned about what word processing software I would use. WordPerfect 4.2 was all the rage, and I was leaning towards that only because it was what was used in the office, until I read glowing reviews in a computer magazine of software I'd never heard of--XyWrite and Nota Bene--and of which I was sceptical, especially because XyWrite seemed complicated . When I received a brochure about Nota Bene from the company, I was sold. Thinking back, I've never used the software for what I thought would be my main purpose, but that, perhaps, has made it all the more useful to me. Besides its myriad and powerful features, especially those geared toward scholarly work, the program made lean ASCII files, which at the time were the preferred means of delivering electronic texts for typesetting. I didn't know what I wanted to publish, but I knew I had the right tool for delivering an electronic manuscript.

One of earliest projects I remember doing was to compose the copy for a small magazine, Shroud Spectrum International, a scholarly journal concerning research of every kind on the Shroud of Turin. I composed all the text using conventional word processing features, so that it could be easily proofread by my colleague--a person not convinced that computers were needed for the job. I offered to input the text as a way to reduce her typesetting charges. When the text was ready, I ran a macro I'd written to change all the formatting codes, replacing them with text strings that the typesetting equipment would use to change fonts and other formatting requirements.

I gradually embarked on several other projects, and since 1990, I have, with my colleague S. T. Joshi, input (by the brute force method) enormous text libraries of the writings and correspondence of Ambrose Bierce, George Sterling, H. P. Lovecraft, Clark Ashton Smith, and others. We did not have a strong plan at the beginning. For example, Joshi had long said that Lovecraft's greatest writing might well be his letters, and yet only his heavily edited Selected Letters had been published. For unknown reasons, I was "inspired" to start transcribing Lovecraft's letter from manuscript. We gradually acquired copies of thousands of Lovecraft's letters over time and input them all. Some volumes of letters have seen publication, but publication of the whole has been slow for any number of reasons. Even so, the vast library of text now input has been useful in all manner of Lovecraft research, thanks to a component of Nota Bene that I thought I would have little use for--that is, the powerful N.B. Orbis tool. One area in which Orbis has been helpful has been in the dating of undated letters. Lovecraft's letters to August Derleth at the State Historical Society of Wisconsin have been grouped by the archivist in such a way that letters without exact dates follow letters with dates (by year) on the microfilm we have. With close acquaintances, or persons with whom letters were exchanged frequently, Lovecraft might date a letter simply "Thursday," but which? Gradually I learned to use the old Nota Bene Textbase, now Orbis, to search through text on various details so as to date letters with precision on the basis of content in letters to other correspondents. There have been cases where we changed the sequence of undated letters for a book, because with Orbis we were better able to ascertain the sequence of events discussed, and then apply a reasonably sound date to a letter.

In the mid-1990s, Joshi, who by the way does not use Nota Bene, wanted to publish a bibliography of Ambrose Bierce, as there was no good bibliography of his work. A journalist for more than 40 years, Bierce wrote thousands of pieces for various newspapers, virtually all of it unsigned. For reasons unclear to me, we undertook not only to find all of Bierce's writings in the newspapers, but also to transcribe all of that work. Including letters, we probably have nearly 5 million words of text in electronic form. Much of what we typed initially helped us to track down Bierce's unsigned work. For example, Bierce's earliest books--The Fiend's Delight, Cobwebs from an Empty Skull, and (to a lesser degree) Nuggets and Dust, published as by "Dod Grile"--contains hundreds of little squibs from his early newspaper writing. We had transcribed all "The Town Crier" columns from the San Francisco Newsletter and California Advertiser, and so I systematically searched for text from every paragraph in those books to identify the source of each item. A surprisingly large number could not be identified at first. Gradually, I developed a technique for searching with Orbis, learning which words to avoid and which to use. Because Bierce often rewrote his newspaper work to suit later publication, he might leave out text, change proper names of individuals, and so on. The list of unidentified pieces gradually grew shorter, but many items were not yet identified. At that point, we went back to the microfilms and re-searched the newspapers. This time around, we found the text to come not from the known Town Crier columns, but instead from various unsigned pieces. Because the text was more or less verbatim in these pieces, we could know with certainty that Bierce wrote them. We also developed a knack for selecting certain items based on writing style and vocabulary, and then went back to the text base searching for text, and found that even in books published in the early twentieth century, Bierce continued to mine his old newspaper writing, even back as far as the late 1860s.

In time, we were able to identify nearly all the original publications of the poems and essays gathered in Bierce's Collected Works. The poems in Black Beetles in Amber and Shapes of Clay were, for the most part, culled right from his newspaper columns, where they initially appeared without titles. A comparison of first lines of verse in the newspapers against those in the published books brought some matches, but a disconcertingly large number of misses. Again, Orbis helped here. I started doing searches in the body of each poem, and started to get more hits in the newspaper columns. We came to realize that Bierce might shorten a poem by removing the opening stanza, now starting it at another line and thus showing why we could not find a match using first lines only. Also, long after the appearance of lines of verse in a column, Bierce might feel that a late book of his satirical poems would have more effect if made more timely, so he would change the names of his victims, or remove or revise now irrelevant information. Again, it took some practice, but gradually we learned what to watch for and what to avoid in doing searches.

Conversely, we were able to make short lists of poems that we hadn't found and go back to the newspaper microfilms looking for them. In many cases, we found the verse in unsigned work that we didn't realize was Bierce's, such as unsigned editorials and articles other than his known "Prattle" columns. Ultimately we were able to deduce the original appearance of nearly all the poems in both books.

One key finding that Orbis helped with was to determine the source of the post-letter R definitions in Bierce's Devil's Dictionary (1911). Bierce prepared his dictionary by culling from published lists of definitions. From the 1880s until the early twentieth century, Bierce on and off published brief lists of his satirical definitions. When the time came to publish his Collected Works, he had worked up definitions only into the letter R, and now needed to write definitions from R through Z to come up with a representative lexicon. The typesetting copy of The Devil's Dictionary exists, and that later part of the book consists of a typescript, whereas the first part of the manuscript consists of a marked-up copy of his published Cynic's Word Book (A-L) of 1906, and late clippings from the newspapers (L-R). We recognized that some of the content in the typescript sounded familiar. Indeed, many of the illustrative verses were recognizable because of their great length. We set about searching through all our typed newspaper columns, using Orbis of course, and found the original sources of virtually all the illustrative verses in the late definitions, and also of many of the definitions themselves. An envelope of various clippings that Bierce thought he might use in writing his dictionary exists, and so we obtained copies of those clippings. Small as they were, we were able to identify the source of many, but we found that others did not exist in our text base. By examining the type styles, we were able to determine what newspapers they likely had come from and so once again consulted the microfilms, now finding the brief snippets to be from other unsigned writings clearly by Bierce.

In preparing Bierce's collected fiction for publication, we were finding that various words were not spelled consistently. How to determine what Bierce's preferred spelling might be? I had one textbase for his published writings and another for his letters. The point was to search each for the words in question, figuring that Bierce's preferred spellings would be evinced in his letters, which were not subject to editing, typesetting, and so on.

Currently Joshi and I are engaged in editing the correspondence of George Sterling and Clark Ashton Smith. Misfiling of manuscripts at a library has separated enclosed poems from the letters that they had accompanied. The ultimate disposition of manuscripts may never be solved, but Orbis has been extremely useful in uniting the majority of poems with their respective letters. Orbis has also been useful in other respects. When either poet mentions an individual by last name only--someone known to the both of them, but not obvious to the reader because perhaps they discussed the person once in person and so need not use a full name in correspondence--I use Orbis to search through other correspondence by both poets to look for clues as to who the person is. When one poet complements the other on a felicitous line of verse, or, more difficult to resolve, suggest a change to a line, I use Orbis to search for the text in question.

I neglected to mention that I am a technical editor at an engineering company by profession and have a full time job--and then some§ So all the research work I do as noted above is "recreational." I have few hours to spend sifting through and memorizing the millions and millions of words that constitute the heart of my research projects, and so Orbis makes me look like a far more intelligent and knowledgeable scholar than I really am. Simply put, Orbis has had a big hand in preparing editions of the writings of Bierce, Lovecraft, Sterling, and Smith that are far more accurate (and more quickly prepared) that would have been possible without it.

David E. Schultz