CIC Wright 19th Century American Fiction Project

Encoding Guidelines

DTD and Files for XMetal, Author/Editor, WordPerfect
Related documents
Introduction
Principles
General Transcription and Encoding Practice
Structural Divisions
Individual Elements and Features
TYPE Attribute Values
Proofreading

DTD and Files for XMetal, Author/Editor, WordPerfect

To retrieve any of these files, hold down right mouse button, and click on "Save Link As" (Netscape) or "Save Target As" (Internet Explorer).

Wright.dtd: DTD for Wright project (modified TEILite)
Wright.xmetal.rls: Rules File for XMetal (change file name to "Wright.rls" after download)
Wright.ae.rls: Rules File for Author/Editor for Windows (change file name to "Wright.rls" after download)
Wright.mac.rls: Rules File for Mac Author/Editor
Wright.wpt: Template File for WordPerfect

Return to top

Introduction

These guidelines are meant to supplement and extend the TEI Text Encoding in Libraries Guidelines for Best Encoding Practices. The texts in this project will be encoded at Level 4 as described in the Guidelines, with more detailed description of encoding practice below.
Return to top

Principles

It is better to do less, than to do wrong or mislead.
All encoding decisions should allow for enhancement and avoid tag abuse.
It should be easy to move from transcription to page images. Elaborate encoding schemes to describe rendition are not necessary.

Return to top

General Transcription and Encoding Practice

One of the primary goals for the project should be accurate transcriptions. By this we mean

Letter-for-letter, not modernized, not expanding abbreviations.
- Accented characters and ligatures will be encoded as entity references. There is a list of ISO Latin and Greek entity references here.
- Quotation marks: Left and right double quotation marks will be transcribed as "; left and right single quotation marks and apostrophes as '.
- Typos or misprints should be encoded as <SIC CORR="">, retaining the original spelling. It can be difficult to determine whether a particular word or phrase is a typo, or just an alternate spelling, or dialect. William Craigie's Dictionary of American English on Historical Principles (Chicago: Univ of Chicago Press, 1936-44) can be useful in determining how a particular word was used or spelled in the 19th century. That said, I try not to go overboard with encoding typos. Try not to encode missing punctuation--I don't encode, for instance, missing periods as <SIC CORR="."></SIC>, but simply add the periods to the electronic text.
Don't transcribe signatures, printer's marks, book plates, running heads, handwritten text
Don't mark line breaks except in verse.
Don't encode features such as foreign words or quotations unless they're somehow highlighted in the text (see Font Shift below)
End-of-line hyphens should be removed when dividing a word at a syllable break, but not for multiple words combined by hyphens. Delete the hyphen, and rejoin the two parts together into one word. It can be tricky to determine, since hyphens are abundantly and inconsistently used in 19th century texts. Try to find another instance of the same word in the same text if it's unclear whether the word should be hyphenated or not.
End-of-page hyphens will be encoded with <ORIG> element, using the REG attribute to record the entire word or phrase, with the same proviso about hyphenated words as described above. Examples:
- <ORIG REG="condition">condi-</ORIG>
  <PB N="5">
  tion.
- <ORIG REG="foot-soldier">foot-</ORIG>
  <PB N="6">
  soldier.

Return to top

Structural Divisions

Title pages, frontispieces, tables of contents, introductions, and prefaces will generally go within the <FRONT> element.
The novels or stories themselves will go within <BODY>.
End notes, publishers' advertisements, and anything else that follows the novels or stories will go within <BACK>.
For most texts, the main text will be encoded within a single <TEXT> element, with multiple numbered <DIV1>s for chapters, sections, stories, or poems.
<DIV>: Use numbered DIVs. Structural divisions within a text can be difficult to identify, and consistency within a work is essential. Most of the novels in this collection should be straightforward, sequentially numbered chapters. Fewer <DIV>s are better than more. Use visual cues as evidence of <DIV>. Strong evidence includes:
- Headings that appear in a table of contents
- Blank page followed by a new heading
- Heading followed by drop cap or ornamental letter
- Numbering scheme associated with headings
Weaker evidence:
- Ornamental device
- Marginal numbered headings
Use other types of chunking if the evidence for a new DIV isn't clear: <P>, <LG>.

Return to top

Individual Elements and Features

Arguments:
Arguments usually summarized the prose or verse that follows. Use <ARGUMENT> only if "argument" is mentioned within the text block. If in question, use <HEAD>.
Drama:
- Cast lists are encoded as a <LIST>. For complex cast lists, use nested lists and labels to indicate cast groupings. Character names, actor names, role description should be included within same <ITEM>.
- Use <SP>, <SPEAKER>.
- Stage directions that occur between speeches or occur within brackets [], and thus easily identifiable, should be encoded as <STAGE>.
Epigraphs:
- These can be tricky to identify. If attributed, epigraphs should be encoded using <CIT><Q> and <BIBL>:
  <EPIGRAPH><CIT><Q><LG TYPE="quotation">
  <L>Since I can do no good because a woman</L>
  <L>Reach constantly at something that is near it.</L>
  </LG></Q><BIBL>--Beaumont and Fletcher.</BIBL></CIT> </EPIGRAPH>
- If unattributed, they should be encoded as just <:Q>:
  <EPIGRAPH><Q><L>Mein Herz, mein Herz is traurig.</L></Q></EPIGRAPH>
- If not sure, they should be encoded as <HEAD TYPE="sub">.
Figures:
- Indicate any illustrations with <FIGURE></FIGURE>.
- <HEAD> within <FIGURE> contains the caption. Any other text associated with the figure should go inside <P> tags. It doesn't have to make sense, but should include dialogue, border text.
- A figure starts with a heading or beginning of a graphic, and ends with a caption or end of graphic. If unclear, it is better to record the caption outside of the <FIGURE>, than to include text not associated with the figure inside of <FIGURE>.
- In the display of the transcription, we recommend to not display text within figure, but instead pointing to the page image. This text would be displayed in any intermediary search results as keywords-in-context.
- <FIGURE> unfortunately is not allowed in many contexts; it generally has to be enclosed with a <P> element. If an illustration occurs, surrounded by a paragraph, you're fine:
  <P>Then our heroine crossed the dangerous
  <PB>
  <FIGURE><HEAD>Our Heroine Crossing the Dangerous River.</HEAD></FIGURE>
  <PB>
  river.</P>
  If an illustration should occur between paragraphs, you must enclose the <FIGURE> element within <P> tags:
  <P>Then our heroine crossed the dangerous river.</P>
  <PB>
  <P><FIGURE><HEAD>Our Heroine Crossing the Dangerous River.</HEAD></FIGURE></P>
  <PB>
  <P>She noticed, however...
  Finally, if an illustration should occur at the beginning of a chapter, you should enclose it within <HEAD> tags:
  <DIV1 TYPE="chapter">
  <PB>
  <HEAD TYPE="figure"><FIGURE><HEAD>Caption [if any]</HEAD></FIGURE></HEAD>
  <HEAD>Chapter II.</HEAD>
  <P>Chapter begins here...
Font shift:
There are many different ways to encode font shifts, depending on their function. Possibilities include:
- If the words are somehow meant to be read with emphasis, use the <EMPH> tag--this is the most common use of italicized text.
- Foreign phrases should be encoded as <FOREIGN>, with the three-letter ISO 639/2 language code as the value of the LANG attribute. Remember, if you include a LANG attribute, you have to declare it in the header in a <LANGUAGE ID=""> element, within the <PROFILEDESC><LANGUSAGE> section.
- Bibliographic titles should be encoded as <TITLE>; names (including, eg. ship names) indicated by font shifts by <NAME>.
- Abbreviations use <ABBR>.
- <HI> should be used in cases where it's unclear exactly what function the italicized or bolded text serves. <HI> should be used for inline italicized quotations.
- Dropcaps, ornamental letters that begin chapters should be encoded as <HI>. Include the entire word within <HI> tags.
Illegible text, gaps, etc.:
- <UNCLEAR> for illegible printed text
- <GAP> where page is torn or not printed.
Lists:
Lists would be defined as those easily identifiable, with repeated categories of data of regular length. Numbered paragraphs would not normally be considered as lists. Tables of contents are lists.
Notes:
Anything printed outside of a text flow. Asterisks, numbers, daggers, double bar etc as N values, and not included in text.
- Use visual cues to group notes. If it looks like one note, it's one note. Notes can appear as indented into text block.
- Text is transcribed until the beginning of the note, then the entire note is transcribed within a <NOTE> element, and then text is resumed. If a note extends over a page, the entire note should be included as a single note. Notes without obvious references are transcribed where they occur.
- PLACE attribute required: inline, margin, foot, end
- <NOTE> will always include <P>
Openings and closings:
- <TRAILER> for the names of publishers or printers that might appear at the very end of a book.
- <CLOSER>THE END.</CLOSER>
- <CLOSER><DATELINE>
- <OPENER><SALUTE>
- <OPENER><DATELINE>
Page breaks:
- The text you receive will have page breaks for each two-page image that precede the pages in the text file.The format will be:
  <PB REF="Wright2-0048-0000023.tif" SEQ="142" RES="600dpi" FMT="TIFF5.0" N="" PAGES="54-55">
  - REF points to the image file;
  - SEQ indicates the sequence of the image, for the page turning software;
  - RES indicates the resolution at which the image was scanned; values will be either 400 or 600dpi.
  - FMT indicates the format of the image; the value will always be "TIFF5.0";
  - PAGES indicates the page numbers and features on the image pointed to by the REF attribute. Features will include "Table of Contents", "Illustration", "List of Illustrations", etc.
- Do not change any of the attributes or attribute values already in the <PB> element. However, if values of the PAGES attribute are in error, correct them.
- Fill in the value of the attribute, N, with the value of the first page on the image. In the example above, the PB tag would look like this:
  <PB REF="Wright2-0048-0000023.tif" SEQ="142" RES="600dpi" FMT="TIFF5.0" N="54" PAGES="54-55">
- At the point at which the second page break on the image occurs, add a new <PB> element, with just an N attribute, e.g. <PB N="55">.
- Page numbers will be recorded as they appear on the page and regardless of where they might occur on the page, as N attribute values in the PB element at the top of the page.
- If there's no page number printed on the page, but the page is actually page 8 in the sequence, put the number within square brackets, e.g. <PB N="[8]">.
- The <PB> element should be placed within any <DIV> tags, and not between. <PB> should be placed between paragraph or stanza breaks, if the paragraph or stanza ends on one page, and a new one begins on the next.
- Blank pages should be marked <PB>, but no further encoding to indicate that they are blank.
Personal letters:
It is common to find personal letters quoted in a text. If a quoted letter includes an opener, dateline, salutation or signature, or is in some other way set apart from the text block, use <Q><TEXT><BODY><DIV1 TYPE="letter">, with <OPENER> and <CLOSER> as appropriate. If a letter is not set apart in some way, but merely appears in the regular text flow, just mark as a paragraph. Examples:
- Example 1:
  <P>He read the letter:
  <Q><TEXT><BODY><DIV1 TYPE="letter">
  <OPENER><DATELINE>Honolulu, Hawaii</DATELINE>
  <SALUTE>Dear Sarah,</SALUTE></OPENER>
  <P>Arrived....</P>
  <CLOSER><SALUTE>Yours sincerely,</SALUTE>
  <SIGNED>John Greswold</SIGNED></CLOSER>
  </DIV1></BODY></TEXT></Q>
  </P>
- Example 2:
  <P>He read her letter:</P>
  <P>"DEAR JOHN: I arrived in Honolulu...</P>
Poetry:
- Each poem is encoded as a numbered <DIV> always using <LG>, even if there is only one line group.
- Happily, most 19th century verse is conventional and regularly printed. Typically, the first letter in each line of verse will be capitalized. This is the visual cue to tell the difference between the beginning of a new line, and line that has been broken. 19th century poetry is metrically very regular too, but since they don't teach prosody any more, students will need the visual cue. If a line looks like this:
  And for you, my noble lady, take my blessing on your
  head,
  it should be encoded as a single line of verse, without the line break:
  <L>And for you, my noble lady, take my blessing on your head,</L>
- When a full line of verse is indented, use the REND attribute, with values "indent(1)", "indent(2)", "indent(3)", "indent(4)", to correspond to the approximate number of indented tabstops:
  <LG TYPE="stanza">
  <L>This is the first line of verse,</L>
  <L REND="indent(1)">This line is indented;</L>
  <L>This is the third line of verse,</L>
  <L REND="indent(1)">This line is also indented.</L>
  </LG>
- Numbered lines will use N attribute.
- Braces that group verse lines (such as triplets) will be disregarded.
Publishers' Advertisements:
Many books contain publishers' advertisements in the back. These will be, for the most part, lists of books. These should be encoded as <LIST>, with each book as an <ITEM>. There is no need to further encode the bibliographic information in these advertisements--the encoding would take far too long if you were to identify and encode authors, titles, publishers, extents, and all of the other bibliographic elements.
Quoted Text:
- Text that is indented as a blockquote should be encoded within <Q> tags.
  - If prose, text should be within a <P> tag even if there is only one paragraph.
  - If poetry, verse should be within an <LG> tag even if there is only one verse block. If there is only one line of quoted verse, then do not use an <LG> tag.
- Inline quoted text that is italicized should be encoded as <HI>.
Tables:
Those texts with important spatial relationships should be encoded as tables; otherwise, the text should be recorded serially in lists or paragraphs.
Tables of Contents:
Tables of contents should be encoded as lists, with the individual items as <ITEM>. Generally, TOC's should go into their own <DIV1 TYPE="contents">. Chapter and poem titles go within the <ITEM>, with the page number within a <REF> element. Example:
<DIV1 TYPE="contents">
<HEAD>Table of Contents.</HEAD>
<LIST>
<ITEM>Chapter 1: Introduction <REF>5</REF></ITEM>
<ITEM>Chapter 2: Conclusion <REF>10</REF></ITEM>
</LIST>
</DIV1>

Similar lists, such as lists of illustrations, should be encoded the same way, in separate <DIV1>.
TEI Header:
The partial TEI Header will be provided for each text, with information taken from the MARC record. This information should be checked against the actual text itself. The Wright number will be included as the ID No. for the text. Use the TEI/MARC recommendations.
Title Pages:
Title pages are structured and formulaic. Once you encode one, you can usually copy it and fill in the fields, with some variation. Some things to note:
- The verso should be included within the TITLEPAGE element, after a <PB N="verso"> tag.
- "By the author of [list of titles]" should go in the BYLINE element
- "All rights reserved" goes into a DOCEDITION element
- DOCDATEs in Roman numerals should use the VALUE attribute for Arabic numbers: <DOCDATE VALUE="1867">MDCCCLVII</DOCDATE>

Return to top

TYPE Attribute Values

DIV
For those structural divisions that can be clearly identified, use the following set of TYPE attribute values in the DIV:
act | advert | book | castlist | chapter | colophon | concluding | contents | corrigenda | dedication | endnotes | epigraph | frontis | index | letter | part | poem | preface | scene | section | undetermined
For those sections called "chapters" use TYPE="chapter". Larger divisions are called "parts" if they have no designation of their own, and that smaller divisions (within chapters, poems, etc.) are called "sections" if they have no designation of their own.
LG
For <LG> within poems, use either TYPE="stanza" (when line groups are regular) or TYPE="para" (for "verse paragraphs, i.e., irregular line groups). For quoted text, where you're unsure if the entire line group is included, use <LG TYPE="quotation">.

Return to top

Proofreading

Proofreading is the most important part of this whole project, and the most difficult. It requires a great deal of concentration and a different kind of reading than you're probably used to. You'll want to record the text as it appears on the page, and be sure that you're taking the time necessary to see the page as it is printed. The best method is generally to read a line of text, read a line on screen. When you find an error, fix it and reread the line to see if there are other errors in the same line. You'll also want to impress upon your staff the importance of accurate proofreading.

I've tried to provide the most accurately OCR'ed pages possible. In fact, I've been OCR'ing most of the pages twice, using different settings, and choosing the more accurate page. But in some cases, the printing is too faint or too dark, or has too much bleed-through, or too many random speckles, or something, and the OCR comes out completely garbled.

The best thing to do in these cases is to retype the page, rather than try to correct the errors in the OCR'ed text--you'll figure out pretty quickly which ones these are. The more errors a page has, the more difficult it is to proofread, because you may fix one error, but miss others in the process. A certain number of titles simply cannot be OCR'ed, and we'll have to consider having these titles retyped by a vendor. Contact me if you think your text has too many errors to proceed.

Some common errors you'll find from the OCR include:

A zero (0) instead of a capital O. You'll want to be sure you can recognize the difference with your software and computer monitor.
Lowercase L, capital I, and the number 1 will all be mixed together. Be sure that you can recognize the difference.
The OCR software may read a capital H as II, or I-I, or somesuch. I'll try to correct the latter (I-I), since that's pretty uncommon, but changing all the II to H would mess up Roman numerals.
A hyphen "-" instead of an em dash "—". Hyphens combine words, while em dashes separate them. Unfortunately, the OCR software doesn't seem able to tell the difference in most cases. You could have the proofreaders use a double hyphen -- for an em dash, and do a global search-and-replace to the entity reference — when they're done (as long as you remember to do it!).
A single quotation mark (') instead of a double quotation mark (") and vice versa.
"h" as "li"; "n" as "ii"; "w" as "vv"; are among the most common mistakes.

Last updated: 26 January 2001
URL: http://www.letrs.indiana.edu/wright/guidelines.html
Comments: PWILLETT@indiana.edu