The CLO has been designed and encoded in XML in order to conform fully to the guidelines set forth by the Text Encoding Initiative Consortium (TEI; www.tei-c.org). Each document has been assigned a registered DOI to establish the means for future connectivity and to ensure the sustainability of the project.

The encoding guidelines for the original CLO were developed in 2001 by Paolo Mangiafico, Perkins Library, Duke University; David Southern, managing editor of the The Collected Letters of Thomas and Jane Welsh Carlyle, Duke University Press; and Brent E. Kinser, at the time an editorial assistant for the project. The initial encoding of the first thirty-two volumes was completed by DNC Data Systems of Mumbai, India
(www.dncdata.com). The first twenty-seven volumes were keyboarded into digital form using a double-entry validation system. Volumes 28–32 were encoded from
extant, electronic typesetting files.

All of the volumes then underwent a thorough two-phased process of proofreading under the direction of the coordinating editor to correct errors both in transcription from the print edition and in the XML encoding of the data. For more detailed histories of the print and electronic editions, follow the links in the CLO footer.

For the new version of CLO the encoded volumes have been updated from TEI P2 to TEI P5, the most recent version available.

Arguably, the most useful justification of editorial process in the print edition occurs in Charles Richard Sanders’s article “A Brief History of the Duke-Edinburgh Edition of the Carlyle Letters”, Studies in Scottish Literature 17 (1982): 1–12. Sanders discusses the collection of the letters (Sanders requested and received thousands of facsimiles from libraries around the world) and the method by which they were organized and stored at Duke:

All the facsimiles and even references to the letters not yet found are listed on 4×6 cards, a card for each letter. . . . We also put on the card all that we know about the letter: where the original letter is if we have found it, where it had been printed if it has, whether or not there are omissions in the printed text, whether we have the letter on microfilm or in a facsimile, and a topical analysis of particularly interesting details in the letter. These cards, filed in chronological order, our guide to all the Carlyle letters that we know about in the world, are kept in a fire-proof filing cabinet. (9–10)

Although much has changed since Sanders wrote this description in 1982, this filing cabinet (no longer fire-proof) remains at the Carlyle Letters Office at Duke University Press, and it also remains a treasured and valuable resource for the current editors. In his article Sanders also discusses the process of editing the letters:

The editorial process with each letter begins at Duke, where the master file is kept. After a typewritten transcript with a carbon copy has been made, the transcript is carefully proofread against the facsimile. It is then ready for the Duke editor to write draft notes. . . . From time to time collections of the carbon copies are sent to the editors in Edinburgh. There the text of the letter is carefully read against the original manuscript of the letter if it is in Edinburgh. The editors then work through the draft-note sheets carefully, altering, adding, and answering queries so far as possible. The notes are then re-typed and returned to Duke with the carbon copies of the text and the draft notes. After the Duke editors have gone through the notes one more time, they are sent to Duke University Press, where Mrs. Joanne Ferguson, Mrs. Myrna Jackson, and sometimes other copy-editors work with the text and notes before they are ready for the printer. (10)

According to Sanders, the editors would then reread the typeset text in galley proof to check for errors and to add as much new material as possible before the beginning of the production phase. Then, as now, new letters and insights were being discovered too late for proper chronological placement in the print volumes.

One of the great editorial advantages of the CLO is that newly discovered materials no longer suffer from being “too late”; they can now appear in their proper chronological place. This change has brought new vigor and relevancy to the editors’ search for misplaced and unknown letters and manuscripts of the Carlyles.

In substance, the editorial process outlined by Sanders remains in place today, and the editors still routinely check the typescripts of the letters against the original manuscripts whenever possible. With the passing of time, however, the order of the processes has shifted.

The American editors, working from the typescripts already checked multiple times according to the process outlined by Sanders, now write preliminary draft-notes that are checked and completed by the editors in Edinburgh.

The Edinburgh editors then also recheck the typescripts against manuscript or facsimile versions (the letters retain their form in typescript in order to prevent the introduction of new errors) and submit them to Duke University Press, along with the electronic files of the notes and the editorial apparatus.

Duke University Press directs the copyediting of the letters, after which they are sent to the Edinburgh editors for penultimate review and then to the typesetters for electronic typesetting and printing.

The American and the Edinburgh editors are then given the opportunity to suggest changes and to correct errors at the proof stage, with the Edinburgh editors directing the final steps to publication. The entire process for each volume takes approximately three years.

Although the source text for the CLO is the Collected Letters, the electronic version does not represent an attempt simply to re-create the print edition. Because of the exigencies of digital searching, for example, silent corrections have been made to incorrect references to other letters and texts in the Collected Letters. Similarly, typographical mistakes clearly not made by the Carlyles, such as missing periods or quotation marks in footnotes, have also been corrected silently.

Other changes to the print edition are documented at the end of the “Sourcenote” in square brackets, as in the case of the letters published too late for chronological placement in the print edition that have been moved in the CLO to their correct date position.

Whenever a change is made to the original text of the print edition, those revisions are retained in the raw XML files through the use of wrappers that are not displayed and in XML comments.

Letters that have been superseded and removed from the collection are identified in the sourcenote of the new version, but users must rely upon the print edition at this stage if they wish to view this superseded material. An example is TC to JAC, 30 May 1826, a letter that was initially published in volume 4 from a published version and in volume 7 from a more definitive manuscript transcription. The definitive manuscript version has been moved to volume 4, its proper place in the chronology of the letters, but unfortunately, at this time the version it superseded is no longer available in the CLO.

The primary organizational principle of the CLO is the date of composition. Every letter is supplied with an XML tag, docDate, which contains the regularizing attribute, “value.” For example, docDate value=”1856-04-30″ identifies 30 April 1856. The letters of the collection are ordered by means of the docDate value.

When the Carlyles do not provide an exact date and when the editors are unable to postulate an exact date via context, docDate has been assigned a value of “00” in the month and/or date position, docDate value=”1856-04-00″ identifies late April 1856. Without benefit of human intuition, the database system automatically reads “00” as “01”; the computer, therefore, reads and orders late April incorrectly as 1 April.

To maintain the sequence established by the editors of the CL, all letters that contain a docDate month and/or day value of 00 that causes incorrect ordering have been assigned exact document dates. These assignments are noted in square brackets at the end of the “Sourcenote.”

The textual notes that began to appear with volume 19 of the print edition as “A Note on the Text” outline philosophical justifications for editorial procedures in the print volumes and continue to evolve over time as conventions change. Because they direct the editorial approach to the copy text for the CLO, the textual notes published for the print volumes remain essentially relevant to the CLO. “A Note on the Text” from volume 32 encapsulates the editorial guidelines for the first thirty-two volumes of the print edition:

Within our editorial conventions we have tried to show as nearly as possible what the Carlyles actually wrote. Commas and periods in TC and JWC’s letters have been regularized within quotation marks. Anything with square brackets that is not noted indicates illegibility, tears, or stains in the manuscript. Damage to manuscripts is normally noted in the headnote. All square brackets by TC or JWC are normalized to round brackets to avoid confusion with editorial square brackets. With TC, who composed carefully, omissions of words and letters and occasional misspellings or other errors are normally indicated by notes. With JWC, who often wrote hurriedly and carelessly, errors are usually unnoted, except when editorial comment is necessary to make the text clear, or confirmation is needed that such and such was what she set down. Titles cited frequently in the notes are abbreviated (see Key to References); others are listed with date and place of publication, with the exception of titles published in London, which give only the date.

In using the Collected Letters as copy text, the text of the CLO adheres to that of the print volumes. Because of differences in medium, the primary editorial goal of the CLO is similar to but different from that of the print edition; it is to display texts that represent a complete and accurate reflection of the content of the Carlyles’ letters as far as the limitations of digital typography and technology allow.

In the CLO, the editorial use of square brackets as defined in the print-edition textual notes has been retained, with a significant addition. Letters in the print edition are introduced in a headnote that begins with the correspondents and dates of the letter, TC-JWC, 4 April. The headnote of the print edition has become the “Sourcenote” in the CLO; it is displayed at the bottom of each letter. In the interest of functionality, correspondents of the Carlyles in the CLO are required to have a unique abbreviation associated with them.

One of the limitations of digital searching is that it cannot distinguish what to the human eye seems logical. Thus John Forster and John Fergus, both JF in the print edition, are now JF and JOFE in the CLO. This change was effected because without unique correspondent abbreviations the web of hyperlinks that connects the letters within the Collected Letters would not function—the computer cannot distinguish between JF (John Forster) and JF (John Fergus).

In the CLO “Sourcenote,” when the CLO correspondent abbreviation differs from that of the print edition, it appears after the print-edition version, in square brackets. In the footnote hyperlinks, however, which automatically take the user to other letters and texts published in the collection, only the CLO abbreviations are displayed, without square brackets, unless there is a question as to the identity of the correspondent (as in the print edition).

Differences in the typographical capability of the analog and digital worlds have changed the way editors and publishers reproduce letters. Although the editorial goal of re-creating as closely as possible the experience of reading the letters in manuscript form remains central to both the digital and the analog projects, one of the strengths of analog typography, the re-creation of special textual instances such as math notation and Greek writing, proves to be a significant weakness of digital display.

Simply put, the encoding tools necessary to make unique textual entities a seamless part of the data flow are not yet available. This weakness, however, is offset by a clear strength, the ability to include actual images of manuscripts within the digital resource easily.

For now, where possible, inline images of special textual instances re-created in the print edition—such as math notation, Greek, and the Carlyles’ descriptive squiggles and drawings—have been included. Clearly, the manuscript version of unique textual entities carries more editorial worth than typographical attempts to re-create them. And initially, users must excuse the rather inelegant manner in which some of the print edition’s typographical entities are displayed. Further, in a few instances, such as the occasional insertion of large braces that enclose addresses in postscripts of letters, the CLO is unable to render a satisfactory display at this time.

Fortunately, such entities are rarely if ever required for scholarly citation; they are, however, of great interest from the standpoint of providing users with an accurate vision of the Carlyles’ creative intent. Additions of “Manuscript” views of the letters that contain such unique textual entities, therefore, will be given top priority as new material is added to the site.

The ODNB after the birth and death dates of figures links to their biography in the Oxford Dictionary of National Biography. Unlike the print edition, in the CLO all references to other letters in the collection use a correspondent and date string (i.e., “see TC to JWC, 29 Sept. 1847”) instead of volume and page numbers. The page numbers, although not displayed, are retained in the raw encoding by means of a wrapper. Each letter and text in the collection contains a bibliographic slugline that appears at the top of the viewing window of letters and other texts. In addition to the DOI, this slugline also identifies the artifact’s location in the print edition.