Your TeX installation must be based on TeXLive 2016 or later (including any recent MiKTeX). If you are on Linux and your distro is something such as Ubuntu 16.04 LTS (Xenial), then the distro version of TeXLive is too old. However, you can still install a more recent version of TeXLive manually, to your home directory, from tug.org.

The novel package is not (yet) included with default TeX distribution. Once you have TeX working, use its package manager to install novel. The TeX package will be placed in (texroot)/tex/lualatex/novel/, the documentation will be placed in (texroot)/doc/lualatex/novel/, and the included font will be placed in (texroot)/fonts/opentype/novel/.

Fonts from the libertinus package are required, and may be installed via your TeX package manager. Note that a novel document does not request the package, so you will not see an error if the fonts are not installed! Instead, if you have not manually specified which fonts to use, it will automatically look for pre-installed Libertinus fonts (or Linux Libertine O) as defaults.

Since you will be using LuaLaTeX, the font cache will be updated to include the newly-installed fonts, the next time you run this software. However, I recommend that you force a full cache refresh, by locating the luatex-cache directory, and discard its contents. The cache may be anywhere (sadly), so you will have to hunt for it. If you have never used LuaLaTeX before, it may not exist. If there are any updates to the fonts, then a manual cache refresh is necessary.

The novel document class, including software, examples, documentation and supporting images, is distributed under the terms of the LaTeX Project Public License, either version 1.3c of this license or (at your option) any later version. The exception is the included font NovelDeco.otf, which is distributed under terms of the SIL Open Font License, v.1.1.

To ensure that your book may be sold commercially, be careful to use only fonts that allow commercial usage. That may exclude some of the fonts in the LaTeX font catalog. See this documentation regarding fonts.

My method of "programming" largely involves using or modifying code provided by others under allowable license terms. Thus, within the TeX code files, you will find places where small or large pieces of code are the work of someone else. Names or organizations are credited in-place within the TeX code, according to license.

Various clipart graphics were obtained from openclipart.org. The NovelDeco.otf font is partly derived from Linux Libertine O, using FontForge.

I read popular fiction a lot, usually books from my local public library. Even if I do not read a particular book, I often look a how it is designed. Throughout this documentation, where I recommend a design or style, it is most often based on the preponderance of designs/styles as I have seen in softcover or hardcover popular fiction, from major publishers in the U.S.A. That does not include "mass-market paperback" versions (the kind seen at supermarkets, rather than at bookstores). I may mention several common alternatives.

I have also consulted the authoritative Chicago Manual of Style (CMoS) hardcover 16th Edition. In the case of fiction, CMoS notes that publishers sometimes don't do things the official way.

Although I have looked at a variety of web sites concerning book design, I give these the least credence, because (a) often they merely duplicate each other's opinions, right or wrong, and (b) they may simply echo CMoS.

In August 2017, an original novel using this document class was published. The text block used nearly all the default settings, and was electronically submitted as PDF/X-1a:2001.

The color cover image was designed in RGB using Inkscape and GIMP, converted to CMYK at 240% ink limit using GraphicsMagick and the methods described in this documentation, then finally converted to PDF/X-1a:2001 using this document class.

Both files were automatically reviewed and accepted on first attempt, with the actual printed books for sale within a few days of file submission.

You must create your document in utf-8 encoding. No OT1, no T1, no LY1, or other TeX font encoding that pre-dates cell phones.

Instead of typing TeX codes such as \`{e} you can simply paste è from a Unicode character map. Other alphabets such as Greek and Cyrillic are directly supported, as long as you choose fonts that support those character sets.

Do not call for fontenc in your Preamble. The novel class is pre-configured for utf-8.

If you use a plain text editor, or TeXWorks, or export plain text from a word processor, be sure that your text is encoded as utf-8, without byte order mark (BOM).

If you are on Windows, do not use Notepad! The free Notepad++ (also Portable version) is a suitable plain text editor. Configure it as specified, with utf-8 and without BOM. Does not matter whether UNIX or DOS line endings, as long as you are consistent.

1.2.1.1. Polyglossia Language Support

The method for loading fonts is described elsewhere on this page. Be sure you read that! For example, you cannot load a font with \usepackage{fontname} in this document class.

With novel, you will be using the LuaLaTeX engine, whether you like it or not. No pdfTeX, no XeTeX, no dvips. You do not have to understand Lua code, since the heavy lifting is already done for you. Just follow the instructions in this documentation.

The built-in fontspec package handles modern Open Type fonts, and is able to access their internal features. The syntax is easy to learn, and very powerful. And, you can directly include text with diacritical marks, Greek, Cyrillic, and most other utf-8 encoded characters in your document. All you need are the fonts, and you probably have them. You will wonder why you never did this before.

Would you like to size text, using command such as \small or \LARGE ? Bah! They will be ignored. Write these commands, and the text will appear at its usual size, just as if the commands did not exist.

There is a reason for this behavior, related to the need for constant line spacing. Thus, novel provides its own capability to size text, using different commands.

Commands such as \section, \chapter, \abstract, \index, \tableofcontents, \maketitle, and many other structural commands will not work with novel. Those are for academic papers. The results they would produce (if enabled) would be inappropriate for fiction. If you attempt to use them, you will either get an error message, or they will simply be ignored.

Novelists do not need many ways to distinguish chapters. With novel there are commands specifically created for the needs of fictional works. They are described below.

Also, the float and table environments are ineffective. In novel there are specific ways to place images or construct a table of contents.

You read that correctly. Sometimes fiction uses images, especially for the title page and chapter titles. With novel, do not use vector formats such as eps, svg, or pdf images. Use raster images in png or jpg format.

There is a reason for this, explained below, where the use of images is discussed in detail. Be sure to read it, as novel has its own way to place images, in a way that is most likely to work well with fiction.

This is standard TeX behavior. Just a reminder!

TeX has a number of special characters. If you type them in a document, they will not appear as an ordinary printed character, but will do something else. These are:

$ % # _ ^ & ~ \ { } and sometimes [ ]

In most cases, an incorrect usage will provide an error message that can be easily deciphered. However, the dollar sign and percent symbol are tricky. Incorrect use of the dollar sign may either place your text in math mode (not your intent) or produce a baffling error message that points to an innocent package. The percent symbol normally produces no error, since it is a comment character. But then, whatever text follows on that line will be missing from your PDF.

Special characters, other than the backslash, are escaped by placing a backslash in front, thus: \$ for the dollar sign, \% for percent. If you must print a backslash, \string\ does it.

• Tip: To avoid unforeseen problems, you should review your document in a text editor, prior to compiling. Search for any instances of special characters, particularly dollar and percent. Backslash them as needed.

You must create your document in utf-8 encoding. You may use a plain text editor (with or without TeX code highlighting), export plain text from a word processor (recommended), or use the minimal TeXWorks GUI. If you use TeXWorks, be sure that you are not running scripts that change the content of your document.

Do not use another TeX editor, such as Lyx or TeXstudio, or an Integrated Development Environment (IDE). Those programs are often capable of inserting standard LaTeX code from menus. However, novel does not necessarily use that standard LaTeX code. The result may be unexpected.

I suggest that you write your novel in a word processor rather than in a text editor. MS Word and similar programs work. I use LibreOffice (also Portable version). If you use LibreOffice, version 5.3 or later is much better than earlier versions.

Why begin with a word processor? These programs can automatically substitute curly quotes and other characters. Remember than since you are using utf-8 encoding, you don't need to use TeX code for that, since you can simply insert the character directly. And, word processors have much better support for spell checking (and sometimes grammar). Just be sure that the word processor does not automatically correct your spelling, because any TeX code won't be in its dictionary.

Remember that TeX closes paragraphs by skipping lines, or by \par. Also remember that in most word processors, the emdash is two hyphens, not the three used in TeX. You can always insert an emdash (or endash) directly. You can write TeX code into the word processor, where it will be treated like ordinary text, so use \textit{text} or similar command, rather than the word processor's italics menu.

In any case, be sure to export plain text as utf-8, then compile it with lualatex.

• Tip: For the editing and proofreading phase, where spelling and grammar are carefully reviewed, I use a special font CPmodified.otf. This is my custom derivative of the Courier Prime regular font. Cpmodified can be found in the documentation extras folder. It is released under the SIL Open Font License v.1.1. To use this font for editing, install it in your operating system, not TeX. Then, call it from your word processor. The font includes Latin-1 and ANSI characters, and a few others. It is not quite monospaced (the em dash and ellipsis are longer). Punctuation is highly exaggerated, so you can see it more easily. My own experience is that most errors in the nearly-finished document involve punctuation.

• Tip: Word processors usually make an incorrect substitution for certain English elisions (possibly in other languages also). When you type "Give 'em hell," the substitution may provide Give ‘em hell, but the correct rendition is Give ’em hell. Can you see the tiny difference in the direction of the single quote? Probably not on your computer screen, but it will be noticeable in print. The CPmodified.otf font makes the difference obvious.

When TeX processes a document on a paragraph-by-paragraph basis, it chooses the layout that has the least total "penalty." Penalties are assessed based on factors such as word spacing, inter-line spacing, hyphenation, widows, and orphans. You may choose the value of each penalty.

With novel, inter-line spacing is fixed, and there is no padding between paragraphs. No plus-or-minus. Hyphens are allowed, but strongly penalized. A reasonable amount of altered word spacing is allowed, and the built-in microtype package allows reasonable changes to letter spacing and the letters themselves. These choices achieve a high level of visual appeal and readability for flowing text.

However, there is little penalty for widows or orphans. When you process a document with novel you are likely to find widows and orphans on many pages. This choice was made for several reasons. First, attention to widows and orphans would reduce the effectiveness of the other, more important, layout penalties. Second, fiction often has numerous short paragraphs (especially with dialogue), and it is very difficult for software to deal with all of them. Third, a great deal of contemporary fiction has widows and orphans. Fourth, it's your book... You can rewrite it!

When you obtain the ISBN for your book, either directly through the issuing agency or via your chosen print service, you will enter ISBN Metadata in an online form. This metadata will go into a database, and identify your printed book in distribution and sales channels. It neither knows nor cares about PDF. (If you were marketing a PDF ebook, instead of print, that would be a different story.)

ISBN Metadata includes title, author(s), subtitle if any, ISBN, other identifying information, and a description. The description is by way of advertising, and can usually be changed later; it need not appear anywhere in your book or PDF file, although some authors might use the description on the book's rear cover or cover flap.

Note: The ISBN refers to your printed book, not to the PDF file. You are not selling copies of the PDF.

Printed Metadata refers to the information that can be seen on paper, in expected places: the book cover, the title page, and the copyright page. Whether your book was printed from a PDF file, or by some other method, is not relevant.

Your print service will insist that the title, subtitle (if any), author, and ISBN in these places exactly match what appears in the ISBN database. Of course, you can (and usually will) use styling or images to make the Printed Metadata more attractive than the plain-text ISBN Metadata.

Some print services may object to artwork that is too fancy, on the grounds that it is difficult to discern the title and author. Although this is very unlikely, it is better to ask before committing resources.

When your PDF file is produced, internal metadata will be automatically created for the benefit of PDF reader software. To avoid confusion, this documentation uses the term "File Data" to mean this internal PDF metadata. You can see some of the File Data when you open the PDF in any reader, and go to File>Properties. To see all of it, you need professional software (such as Adobe Acrobat Pro).

Be sure to read the rules for File Data, below. Some of it is generic, and some is particular to TeX or to the novel class.

If PDF/X compliance is required, then you must at least set the title in File Data, using the \SetTitle command. Note that this title is not styled in any way.

Some print services require a minimum amount of file data, particularly title and author, which must match the printed metadata and ISBN metadata for these fields. Other services may require that the book's ISBN number, with something like _text added, be used in place of the title (remember that underscore is a special character in LaTeX, so it must be preceded by a backslash). You will have to communicate with your service, and determine their requirements.

Many print services do not care about File Data, one way or the other. Your PDF is passed through production according to its file name, not its internals. In such a case, assume that title and author are required.

Caution: Be sure that any included images are stripped of their own image metadata. The procedure is described on farther along in this documentation.

At least in the USA, your print service may require that File Data be limited to characters in the Latin-1 set. This is because their production and distribution software uses Latin-1 for the database. Note that Latin-1 does not include curly quotes or em dash. You will still be using utf-8 encoding in your document files, but your choice of characters is restricted.

If it were not for the database limitation to Latin-1, you could give your book a Greek or Cyrillic title, because novel can do it, and so can the PDF.

If your book has File Data using quotes, then use the special commands \straightquote{} and \straightdblquote{}. Be sure to use the braces, so that you don't have to worry about gobbled space. When you use these commands instead of typing ' or " you eliminate the risk that your software will automatically convert them to the forbidden curly quotes. Example title:

\SetTitle{That\straightquote{}s Amore} % That's Amore

If you are limited to Latin-1 in File Data, you can still use curly quotes and em dash on the cover and title page, and in headings. In novel there is no \maketitle command, so you may (and should) apply styling to the printed title, even though the File Data title cannot be styled.

In File Data, you must obey the same TeX rules that apply everywhere: Certain characters have special meaning, and must be preceded by a backslash, if you wish to use them as ordinary characters. These are:

# $ % & _ { }     (hash, dollar, percent, ampersand, underscore, left and right curly brackets)

Example: The title Doing 9% & Gone is coded:

\SetTitle{Doing 9\% \& Gone}

The above characters are rarely needed in File Data. Sometimes they must be spelled out, in order to agree with your book's ISBN Metadata. For example, you may need to write "and" rather than the ampersand.

• Do not use the backslash \ except as described above.

• Do not use the tilde ~ or the circumflex ^.

• Do not use TeX code, such as \`e for e with grave; instead paste è directly from a character map.

• Do not use styling, such as {\small text} or \textit{text}.

• Use only \straightquote{} and \straightdblquote{} if necessary, not TeX code for apostrophe or quote.

• Do not use multiple hyphens to create longer dashes. Do not use em dash or en dash.

If you do not use the \SetTitle{} command, or leave it empty or blank, then the resulting PDF cannot comply with PDF/X standards. In very rare cases, that is desirable.

This information is not automatically transferred to your book's title page. It will be used as default in certain page heading styles, but you can over-ride it. The information can be retrieved as \theTitle anywhere in your document.

This information is not automatically transferred to your book's title page. It will be used as default in certain page heading styles, but you can over-ride it. The information can be retrieved as \theAuthor anywhere in your document.

Use the name as it is spoken. Correct: Mary Smith. Incorrect: Smith, Mary. Possibly correct, if true: Mary Smith, Ph. D. (ask your print service).

There is no file data for subtitle! The only reason for using this command is if you wish to later retrieve the information as \theSubtitle in your document.

The Application, also known as CreatorTool, is the program used to create your source document. Text editors, word processors, and page layout programs are the most common Applications (not specific to TeX). The PDF Producer is the program that converts the source document to PDF. In many cases, the Producer is a back-end, directly linked to the Application.

These details are automatically created when you compile your document, without any input from you. However, if you have the need to provide a custom text string for the Application or Producer, these commands will do it.

The following choices are available:

X-1a:2001

X-1a:2003

X-3:2002

X-3:2003

off (default)

The first of these is the one that is most widely accepted in commercial printing. Anything meeting X-1a:2001 necessarily meets the other standards. If your book involved complicated color artwork, then the correct choice might be crucial. But for a black and white book block, X-1a:2001 is the preferred choice unless your print service says otherwise.

Unless compliance is off, an Output Intent is always required. However, novel treats this as an optional argument, because it will automatically use CGATSTR001 as default Output Intent.

Output Intent informs the printer that you "intend" your PDF to be printed according to certain printer capabilities. If you took the trouble to soft-proof your PDF on a color-calibrated monitor using color-managed software, and if the printer's capabilities actually correspond to the intent, then the printed result will closely match what you expected. This is very important in commercial color printing (such as for cover artwork), but nearly irrelevant for a black and white book block. Nevertheless, an Output Intent is required to meet PDF/X compliance.

There are three pre-defined choices:

CGATSTR001 (default)

FOGRA39

JC200103

The first of these, more properly named "CGATS TR 001" with spaces, is also known as "US Web Coated SWOP v2" in the USA, and is the most commonly-used American print standard for this kind of book. The second is widely used in Europe, the third in Japan.

If you are required to use an Output Intent that is not one of the above three, then you have to code it yourself. Have a look at the three pre-configured Output Intents (files novel-CGATSTR001.clo, novel-FOGRA39.clo, and novel-JC200103.clo). Instructions are inside the files. There are several places where you can find the necessary data; the most comprehensive is at the International Color Consortium.

In essence: If you create a custom Output Intent file named novel-MyFunkyPrinter.clo, then you request MyFunkyPrinter (without the preceding novel- or the .clo extension) as Output Intent, and that file will be loaded.

Each Output Intent has a corresponding color profile, in the form of a file with extension *.icc.

You usually do not need it. In fact, unless your print service specifically asks you to embed the file, don't do it. When you use \SetPDFX in its un-starred form, the *.icc file will not be embedded.

If your print service demands that you embed the *.icc color profile, then use \SetPDFX* (with the asterisk).

Repeat: Do not embed the *.icc color profile unless specifically requested by your print service.

• If your printer requests "US Web Coated SWOP v2" and, as usual, requests that you not embed the corresponding *.icc color profile, then some software will read the Output Intent as "CGATS TR 001" instead of "US Web Coated SWOP v2". They are identical. But not everyone knows that.

You do not need an *.icc color profile unless you choose to embed it.

Although many *.icc files are freely available, they usually cannot be distributed within an open-source project. You might already have them in your system, wherever such files are stored. Note that TeX uses forward slash in file paths, not backslash, even if the platform is Windows:

Windows: C:/Windows/System32/spool/drivers/color/   [older Windows: system32 instead of System32]
Linux: /usr/share/color or hidden ~/.local/share/icc/ or other places
OSX: /Library/Colorsync/Profiles/ or hidden (username)/Library/Colorsync/Profiles/

If you do not have the files, look on the Internet here and here. These links also have many other profiles, beyond the basic three. In a few cases, your print service may provide the necessary file.

You want printer profiles, usually CMYK. Don't worry about the files being "for Windows," as they are actually cross-platform.

If your system has the necessary *.icc file for your Output Intent, but novel cannot find it, then copy the file to the same folder as your TeX document.

Again: Do not embed the color profile unless the printer requests it.

Trim Size is the finished size of your book, width and height (but not thickness). In the case of softcover books, it is the exact size. In the case of hardcover books, where the cover is larger than the contents, it is (usually) the cover size that is measured, with the internal trim size being slightly smaller than nominal (some of the margins shaved away).

Your print service has a list of standard trim sizes. For fiction, you probably want creme (off-white) paper, which eliminates a few of the choices. Choice of size may also be limited by the distribution channels you choose.

Trim Size is set using this command in Preamble:

\SetTrimSize{width}{height}

The width and height must have dimensions, such as in or cm.

If you do not specify the Trim Size, the default is 5.5in wide, 8.5in high. This is a widely-used "trade" size for softcover print-on-demand fiction in the U.S.A. If in doubt, leave the default.

Do not request a mass-market paperback Trim Size unless you really, truly, know what you are doing. Since you are not a major writer for a major publishing house, you don't know; so, don't do it.

In PDF terminology, Media Size is the size of the PDF page. By default, it is identical to the Trim Size.

Some print services insist that the Trim Size be floated in a larger Media Size, which is likely to be U.S. Letter or A4 paper. This quietly happens when you print at home: The printout is at tray paper size, whether you like it or not.

This is illustrated at right, where the Trim Size (green border) is centered in a larger PDF Media Size. The green border, known as the TrimBox is invisible, and does not print. Anything outside the green border does not appear when the book is finished.

If you are required to float the Trim Size in a larger Media Size, here is the command in Preamble:

\SetMediaSize[alignment]{width}{height}

If the optional alignment is set to edge then the Trim Size will be placed at the edge of the Media Size, rather than vertically and horizontally centered.

Again: You usually will not set Media Size manually.

This command is only effective if you use the cropmarks class option, as described in section 2.4. Note that user-provided cropmarks are very unusual in the print-on demand market. Do not use them unless your print service specifically requests them. The built-in cropmark capability is very elementary, and cannot be easily enhanced by adding other LaTeX packages.

In the accompanying image with cropmarks, a corner of the Trim Size is at lower right. The invisible TrimBox has been drawn with a green line. The cropmark gap is G.

Default gap is 0.125in.

The novel class has its own way of specifying the main font. You do not write \usepackage{fontname} or \RequirePackage{fontname}. If you do that, then a warning or error will result, and in any case the font you get won't be the one you were expecting.

Choose only Open Type or TrueType fonts, not the ancient Type 1 (PostScript) or Metafonts. The font you choose does not have to be part of any TeX package, because you do not need any *.sty file or TeX metrics. All of that is built into the font itself. This is thanks to LuaLaTeX and the built-in fontspec package.

4.2.1.1. \SetParentFont[features]{fontname}

4.2.1.2. \SetFontSize{length}

The descendant fonts are automatically set, as variations of the parent font:

For example, the main textfont uses the parent font family. It has TeX Ligatures and Kerning (inherited from the parent font), and also Common Ligatures and OldStyle numbers. The main textfont is not configured directly. Instead, configure the parentfont, then the textfont (and other descendants) will obey. For example, if you don't want OldStyle numbers:

\SetParentFont[Numbers=Lining]{fontname}

Then, Lining numbers will be used for the textfont (and other descendants).

If you don't like the defaults, you can address the other descendant fonts directly, in Preamble:

\SetChapterFont[features]{fontname}   % see section 5.1.3.1
\SetSubchFont[features]{fontname}   % see section 5.1.3.1
\SetHeadFont[features]{fontname}   % see section 4.3.4.2

You may use \parentfontfeatures and/or \parentfontname as arguments, if you wish.

Open Type font NovelDeco.otf is packaged with the novel class. It is a special-purpose font that provides characters for the \decoglyph and \acronym commands.

Normally, you do not need to configure this font, as it is automatic. The only reason for changing it is if you are an advanced user, who wishes to use a substitute font with different decorations. However, you may not specify just any font! The substitute must be an edited version of NovelDeco.otf, with a different font name. If you don't know what that means, then don't do it.

\SetDecoFont[features]{fontname}   % see section 5.1.3.3

There are several pre-defined fonts that do not descend from the parentfont. In each case, an appropriate default font is selected. You will probably never need to use them; but TeX being what it is, default fonts are chosen "whether you like it or not."

The default sans-serif font will be Libertinus Sans if available. Second choice is Linux Biolinum O. Third choice is Latin Modern Sans. To choose your own, use this command in Preamble, with fontspec syntax:

\SetSansFont[features]{fontname}

The default monospaced font will be Libertinus Mono if available. Second choice is Linux Libertine Mono O. Third choice is Latin Modern Mono. To choose your own, use this command in Preamble, with fontspec syntax:

\SetMonoFont[features]{fontname}

If you need a math font, then may I gently suggest that maybe novel is not the right document class for you? Nevertheless, the (hopefully never used) math font will be Libertinus Math, or Asana-Math, or Latin Modern Math. You can choose your own in Preamble, using the built-in unicode-math package notation:

\SetMathFont[features]{fontname}

You can define new font commands in Preamble. The syntax depends on whether the new font is part of a family (where you might have bold and italic variants), or just a single file (no variants). If part of a family, use \NewFontFamily. If alone, use \NewFontFace.

The \NewFontFamily command (or \NewFontFace) takes three arguments: First, the name you assign to the new font command is required. Second, font features are optional. Third, the font name is required. The fontspec syntax is used:

\NewFontFamily\pickaname[features]{fontfamily}
\NewFontFace\pickaname[features]{singlefont}

Note that if you define a font family when you should have just defined a font face, there will be complaints in the log file, because the font loaded cannot find the (nonexistent) other members of the family. But the document will compile. On the other hand, if you define a font face when you should have defined a font family, then commands for bold or italic will have no effect, since the other family members were not loaded. Again, the document will compile.

Example: Suppose your work is science fiction, and one if its characters is a talking robot. You wish to depict its utterances in a robotic-looking font, which will be called using a new \robovoice command. You shop around, and find that the Orbitron font suits your purpose, except that it is too large compared to your textfont. So, you scale it.

In the Preamble, you define the new font command:

\NewFontFace\robovoice[Scale=0.8]{Orbitron} % not part of a family

After some experimentation, you realize that the \robovoice letters need to be spread apart a bit. Do not use the standard LaTeX \textls command. Instead, go back and change the font definition like this:

\NewFontFace\robovoice[Scale=0.8,LetterSpace=number]{Orbitron} % pick a number

In the document body:

``Want some oil?'' asked Captain Withens.\par
{\robovoice ``Is it Gamma-42 Special oil?''} asked the robot.\par
``No, I got it on sale at Cosmo-Mart.''\par
{\robovoice ``Forget it. I'm an oil gourmet.''}\par

The result:

Caution: If you are using a new font within body text (as above), then it is a bad idea to use Scale greater than 1.0. Reason: An excessively tall font will cause lines to spread apart.

Example: Suppose that you wish to style the title of your book, in a font that is not used for other things such as chapter titles. You purchase a license for the Bernhard Modern Std fonts, in Regular and Italic.

In the Preamble:

\NewFontFamily\booktitlefont{Bernhard Modern Std}

The title page, using \charscale for size:

\thispagestyle{empty}
\vspace*{7\nbs}
\begin{center}
\charscale[3]{\booktitlefont The \textit{Fluffy} Conspiracy}\par
\vspace{6\nbs}
\charscale[2]{Priscilla Milde}\par
\vfill
Nonexistent Publications • Portland, Oregon\par
\end{center}

Margins are required on all four edges of the page. They do not have to be the same width. Generally the margin at the inside edge (spine margin) is wider. The novel class does not provide for marginal notes.

In the case of fiction, it is usually the case that the inside margin is the widest. But this is not a rule for all books. All that is required is that each margin be "wide enough" for the printing process.

Keep in mind that print services are not used to seeing books with exotic margins (such as bottom margin twice the size of top margin). They may think that there is a mistake in your PDF, or in the print job, if you do that.

Normally, your TeX document will be compiled to a PDF that is electronically at the exact Trim Size. And normally, print services want it to be that way. When you view your PDF, the text will shift from side to side as you move through the pages, due to the slightly increased margin at the inside edge.

Margins are always relative to the Trim Size. If you are using a larger Media Size, the area outside the TrimBox is not counted among the margins; it is simply wasted area. This is illustrated at right, where a smaller Trim Size is floated in a marger Media Size. The margins are shown in gray.

Here is how margins are set in Preamble:

\SetMargins{top}{outside}{bottom}{inside}

Mnemonic for order of margins: "TOBI, or not TOBI, that is the question."

Each of the settings must have dimension, such as in or mm.

If you do not use this command, then margins are pre-configured. The minimum pre-configured margins are 0.5in at top, outside, and bootm, and 0.75in at inside. This meets the requirements of nearly any printing technology, unless your book is extremely thick. If your Trim Size is among the larger ones, then the pre-configured margins are larger than these minimums.

If your pages have header/footer, they sit inside the rectangle bounded by the four margins. That is, they do not sit inside the marginal areas.

4.3.1.1. Gutter? Which Gutter?

Lines per page refers only to the textblock, and does not include header/footer. Lines are not stretched to fill short pages.

By default, a suitable number of lines per page are set, based on other dimensions. You can change the default with this command, in Preamble:

\SetLinesPerPage{integer}

Any non-integer will be rounded. There is no direct command to set the leading, also known as line-to-line spacing or baselineskip. Instead, it is calculated from the textblock height and number of lines.

If the calculated leading is less than 1.2x the font size, an error will result. It is difficult for TeX to honor all layout settings with such a small leading, which would normally be unacceptable for fiction in any case. If the calculated leading is at least 1.2x the font size, but less than 1.25x the font size, then the log file will have an Alert message, suggesting that you should consider using fewer lines per page. If the leading exceeds 1.4x the font size, then the log will have an Alert, suggesting more lines per page. The Alert messages are friendly; you do not necessarily have to do anything.

Global choice of header/footer style is made in Preamble. Although some individual pages may have no visible header or footer, in reality the header/footer is "still there" in terms of occupying space, but has no content.

There are six pre-configured header/footer styles, numbered 1 through 6. These address every style I have seen in fiction. When you choose one of these styles, it reserves space for a header (if any) above the main text, and space for a footer (if any) below the main text. You may not write multi-line headers or footers.

The default style is 1, which has header, but no footer. Page numbers appear at the outside. Text (such as author or title) appears centered.

Style 0 has no header or footer. This is not used in fiction, but it is available.

Throughout the body of your document, you can use local commands to change the content of headers and footers, or blank them. But you may not change the overall header/footer style.

4.3.3.1. List of Header/Footer Style Choices

 

\SetHeadFootStyle{1}
Only Header.
Page number at outside (left verso, right recto).
Optional emblem adjacent to page number.
Text centered. Default author verso, title recto.
This is the default for the novel document class.

\SetHeadFootStyle{2}
Only Footer.
Page number at outside (left verso, right recto).
Optional emblem adjacent to page number.

\SetHeadFootStyle{3}
Only Footer.
Page number centered.
Disregards emblem, if coded.

\SetHeadFootStyle{4}
Only Header.
Page number at outside (left verso, right recto).
Optional emblem adjacent to page number.
Text towards outside, instead of centered.
Text begins or ends 1em from the emblem.
Default author verso, title recto.

\SetHeadFootStyle{5}
Header and Footer.
Page number centered in footer.
Disregards emblem, if coded.
Text centered in header.
Default author verso, title recto.

\SetHeadFootStyle{6}
Only Header.
Page number at outside (left verso, right recto).
Optional emblem adjacent to page number.
Text towards inside, instead of centered.
Default author verso, title recto.

4.3.3.2. Custom Header/Footer Style

You may customize the appearance, and to some extent the content, of the various parts of the header and footer. These are global settings, in Preamble.

4.3.4.1 \SetHeadJump{number}, \SetFootJump{number}

4.3.4.2. \SetHeadFont[features]{font}

4.3.4.3. \SetLooseHead{number}

4.3.4.4. \SetPageNumberStyle{code using \thepage}

4.3.4.5. \SetEmblems{verso}{recto}

4.3.4.6. Setting and Changing Header Text

There are a number of cases where a specific page must have a header/footer style that appears to differ from the global style. The key words are "appears to differ." The global header/footer is always there, in terms of vertical space. But either or both may be blank, or have unusual content, on certain pages.

4.3.5.1. \thispagestyle{choice}

4.3.5.2. \SetChapterStartStyle{choice}

Unless your chapters run continuously without page break, begin a new chapter using \clearpage or \cleartorecto. The first of these provides a new page, which will be verso or recto, without skipping a page. The second forces the chapter to begin on a recto page, if necessary inserting a blank verso.

Note that the common modern practice in fiction is to allow new chapters to begin either verso or recto, once you get past Chapter One. The only recent book I've read (from a major publisher), where chapters always started recto, had about 20 blank pages throughout the story. I believe the reason was to pad the book, so that it looked bigger. If you are using an American P.O.D. service, such padding might knock about $0.10 (estimated) from your profit, for each copy sold. So it's up to you to think about the tradeoffs between style, profit per book, and expected sales.

Prologue (if any), Chapter One, Epilogue (if any), and other special sections always begin recto.

Don't worry about using \clearpage or \cleartorecto twice. The compiler is smart enough to understand that you only meant to do it once, as long as there is no other code in between.

The ChapterStart environment creates a block of white space, with a fixed height (number of normal lines). If you do not place much inside it, or even if you place nothing inside it, the space is the same.

If you place too much inside it, the space does not grow. Instead, the excess material overflows the bottom of the environment. There, the excess will underlie the main text. This is usually undesirable, but can sometimes be used for special effects.

5.1.2.1. \SetChapterStartStyle{choice}

5.1.2.2. \SetChapterStartHeight{lines}

5.1.2.3. \begin{ChapterStart}[lines][style]

You can put anything you like within ChapterStart, including ordinary text and even images. All ChapterStart does is reserve a fixed space. However, there are special commands that make it easy to place a chapter title, chapter subtitle, and decorative element.

5.1.3.1. \SetChapterFont[features]{font}
5.1.3.1   \SetSubchFont[features]{font}

5.1.3.2. \ChapterTitle[alignment]{text}
5.1.3.2.   \ChapterSubtitle[alignment]{text}

5.1.3.3. \ChapterDeco[alignment scale]{text}

5.1.3.4. Chapter Epigraph

You can use an image in chapter titles. If you place an image, use one of the image commands defined by novel elsewhere in the documentation. Any other method is likely to push the following text off-grid.

Be sure that you are licensed to use the image. Many clipart catalogs "for personal use only," prohibit usage in commercial books. One possible source is openclipart.org (note the .org).

5.1.4.1. Text Over Image

5.1.4.2. Text Built Into Image

Since the ChapterStart environment does not automatically begin a new page, you can use it for continuously running chapters. In Preamble, set a short height, such as \SetChapterStartHeight{4}. Also use \SetChapterStartStyle{fancy} so that headers don't disappear.

Throughout the book, you can make local adjustments to the ChapterStart height, so that they don't fall at an inconvenient location (such as at the bottom of a page). Yes, I have seen this done in published works. If a particular chapter is sufficiently important that it deserves a page break, or a lot more space, you can do that easily.

The \QuickChapter command inserts two blank lines. In the gap, the chapter title is left-justified using subchfont (not chapterfont). It is slightly raised from the normal baseline, because this looks better; but the following text is back on-grid. By default, the start of the following text will be un-indented.

If the optional linelength is provided, then a dash will be written after the title, at the position and thickness typical of the em dash in subchfont. If linelength is specified in units of length, then that will be the length of the line. But if linelength is an asterisk, then the line will fill to the right margin.

Note that this command is probably too obtrusive for mere scene changes within a chapter.

Example:

Then he looked at me with a sneer, and barked, “You better haul your can over to Vinny's joint, before the Boss finds out.”\par
\QuickChapter[3em]{16. At Vinny’s}
So I took his advice, and hauled my can over to Vinny’s. It was everything I had expected, and worse.\par
The bouncer wasn't the biggest one I’d ever seen, but he was the ugliest. ...

This global command may only be used in the Preamble. Default is false.

Normally, text begins without a paragraph indent, after each scene break. This is standard in fiction. If you prefer that each new scene begins indented, use this command. It works only with \scenebreak, \sceneline, and \scenestars.

Note: \SetScenebreakIndent is new in version 1.40.3. When set to true it is identical to the \IndentAfterScenebreak command from prior versions. The older command is still honored. The new command has syntax more consistent with other novel settings.

Use the \charscale command to size the first letter (or more than one letter) so that it stands above the remainder of the line. Remember that you will see a capital letter standing next to lowercase letters, so it will appear taller. You may also use a font command. Example, where \NewFontFace\alexbrush{AlexBrush-Regular.ttf} appears in Preamble:

\noindent\charscale[1.6]{\alexbrush I}\hspace{0.2em}t was a dark and stormy night. Well, it really wasn't all that dark, seeing as how this story is set in Las Vegas.

If you want to place a portion of the first line in small caps, then simply do it using \textsc{text}. This effect can also continue to additional lines. That is, it does not automatically stop at the end of the first line.

The magaz package is pre-loaded in novel. It provides the \FirstLine command, which takes the entire first paragraph as its argument. Place \noindent inside. You may also style the first letter (or more than one) as described above; the styled letter appears inside the \FirstLine command. Example:

\FirstLine{\noindent\charscale[1.6]{\alexbrush I}\hspace{0.2em}t was a dark and stormy night. Well, it really wasn't all that dark, seeing as how this story is set in Las Vegas.}

The lettrine package is pre-loaded and pre-configured in novel. The configuration forces lines to remain on grid. Lines are not added if the drop cap rises above prevailing text (DiscardVskip is very large). The Findent is set to 0.25em, and Nindent is set to 0em.

For information, see the lettrine package documentation and examples. Be aware that creating good-looking drop caps is harder than you think, even with automation. That is because you will be visually comparing an enlarged letter to surrounding normal text, and it is difficult to align the drop cap with the margin, in a way that looks right. In any case, drop caps are rarely used in popular fiction.

Do not use the [image=true] option with lettrine. Instead, place an image using novel's own \InlineImage command. If you set its vertical offset at baseline, then the bottom of the image will align with the baseline of the final line in the displaced text.

Note that \lettrine has an additional argument after the drop cap.

You cannot use \lettrine in conjunction with \FirstLine. This is a fact of life. If you wish to use a drop cap and also have the first line in small caps, do this: Use empty braces {} as the final argument to \lettrine. Then use \textsc with enough text to continue to the second line. Note where the first line ends. Then, move the closing brace of \FirstLine to that position. You may place the closing brace in the middle of a word, at a hyphenation point, if you wish (do not manually add the hyphen). This is the only method that provides a satisfactory layout.

Ah, the simplicity of TeX font size commands, such as \small and \LARGE. But in novel, they don't work. If you write {\LARGE text}, then the included text appears at normal size. This is by design.

Why? In novel, the focus is on maintaining a constant baselineskip, so that facing pages have text aligned. This is impossible with the standard TeX commands. In particular, commands to increase font size will change the baselineskip, whether you like it or not, even if you try to code for a constant baselineskip. That's just the way TeX works.

The one exception is that footnotes will appear with smaller than normal size and skip. But this is done automatically. You cannot manipulate it with a size command.

Fortunately, there are not a lot of occasions where fiction needs to change font size. In novel, there are new commands to address each situation.

The \charscale command may be used for a short amount of text, where there is no line break. The command will throw an error if the included text has \\ or \par or any other command that breaks the line. Text within \charscale does not wrap. You may not place an image in \charscale.

When used without the option, \charscale echoes the included text. However, the text gains the magical ability to extend above or below the normal line of text, without pushing apart the baselines. This is useful if the text has some unusual character that is extraordinarily tall or deep.

Using the option, you may uniformly scale the text by any amount, and displace it horizontally and vertically. The text can overlie other text, or any previously-placed image.

With the unstarred command, the scaled width of the text will occupy horizontal space, even if the text is displaced far away. But with the starred command, the text does not occupy any horizontal space.

The single optional argument may include up to three comma-separated values. The first value is a number, indicating the scale. If a second argument is used, it is the length of horizontal offset (positive is right). If a third number is used, it is the length of the vertical offset (positive is up). Remember that if all you want to do is scale the text, without positioning it, then you only need to provide the scale.

• Note: \charscale does not commute with some other TeX commands. That is, the results of \charscale{\othercommand{text}} and \othercommand{\charscale{text}} may differ. That is just the way things are. Use the one that works for you.

Here is an example illustrating the command in all its glory:

Bada \charscale*[1.25,5em,1.3\nbs]{Boom} Bam\par
Bada Bam\par
Bada\charscale[1.25,5em,1.3\nbs]{Bang} Bam\par
Bada \charscale[1.25]{Bang} Bam\par

In the result (above right), notice how the starred command leaves no gap where the word Boom would have been. The unstarred command leaves a gap where the word Bang would have been located. In each case, the baseline grid is maintained.

The most practical application of \charscale is styling the first character of chapter text. Drop capitals are out of fashion, but enlarged capitals (rising above the baseline) are in fashion. Example, also using the \FirstLine command (from package magaz, included with novel:

\FirstLine{\noindent\charscale[1.5]{I}t was a dark and stormy night. Lord Withens mounted his horse and rode through the gloom to Castle McDingle, where Baron von Feathering awaited. The pair had dirty deeds to do, but being members of the nobility, they commanded a high price.}

In the above example, the initial I is 1.5x its normal size. It appears larger than that, by comparison to the adjacent small caps letters.

The parascale environment is used on a paragraph basis. More than one paragraph may be included. The scale is applied to the font size and to the baselineskip within.

Since the text within parascale generally will not occupy an integral number of normal baselineskips, the following text will be off-grid. This may not be a problem, if parascale is used in places where maintaining the grid does not matter (such as on a title page or copyright page).

• If parascale is used within body text, then it will be necessary to restore the line grid for the following text. This cannot be done automatically; you will have to make a preliminary run, look at the results, and adjust accordingly.

For example, suppose you use \begin{parascale}[0.92]...text...\end{parascale}. The preliminary run shows that the included text occupies 7 lines. The following baseline will then be 7x(1–0.92)=0.56 normal baselineskip too high. This can be fixed by adding \vspace{0.56\nbs}. For best appearance, the \vspace should be split above and below the parascale environment, but not nessarily in half.

• If parascale crosses to the next page, then you will need to make vertical adjustment for each page independently. For example, if a 7-line parascale has 3 lines at the bottom of a page, followed by 4 lines on the following page, then you need to restore only a 4 line deficit on the following page.

• Within parascale you may need to manually indent paragraphs, or remove indents if necessary.

If you will be using a font in a way that it will always be at a size different from the normal font size, you may use the Scale feature when you first define the font. This is defined by the fontspec package, which is automatically loaded with novel. Example:

\NewFontFamily\minifont[Scale=0.75,Numbers=Lining]{\parentfont}

A scaled font can be used anywhere. It is not limited to a single line, and it will maintain the line grid at scales not exceeding 1. However, scaled text may disturb the line grid if its scale is more than 1.

\textit{text} puts the text in italics, provided that the font family has an Italic font. If it does not have an Italic font, then the letters will remain normal, not fake slanted.

If the italicized text leans too far into the following space, try adding \/ (back and forward slashes) just before closing the braces.

\emph{text} changes regular text to italics, and italic text to regular. You may nest \emph commands. Many writers opine that this is the preferred way to specify Italics.

Sadly, the above two commands will fail if the styled text is in more than one paragraph. A more general command that is able to cross paragraphs is \itshape{text}.

\textbf{text} puts the text in whatever font you specified as the Bold font, when you defined the currently-used font family. If the text is in more than one paragraph, use \bfshape{text} instead.

If the font family has no Bold font, then it will not be faked.

The default main font used by novel, namely Libertinus Serif, has both Bold and Semibold weights. Given this choice, novel explicitly calls for the Semibold weight. If you intend to use a font such as Adobe Garamond Pro, which also has a Semibold weight, then you can request it instead of Bold when you define the font family, using fontspec syntax.

What if you want both Bold and Semibold, or maybe even a Light weight? Only one can be assigned as the bold weight when you define the font. For the others, use \NewFontFace to assign the desired weight to its own font command, which you will use (but inside the braces) instead of \textbf.

Actually, you should not be using Bold in fiction. It might be the case that this weight is useful for titles, but you ought to consider a fancy-looking font for that purpose, rather than mere Bold. Or, make the letters big and bold, but balance the effect by using gray.

Underlining is inappropriate for fiction. Don't do it. Besides, there are some issues regarding line grid and appearance, which are not a problem in academic works, but are noticeable in fiction.

If you are using a line for its graphic effect, the \bigemdash command might be helpful.

\textsc{text} converts lowercase letters to small caps, leaving uppercase unchanged. The font's smcp Open Type feature will be used. But if the font does not have smcp, it will not be faked. If you need to corss paragraphs, use \scshape{text}.

\smcp{text} is an alias for \textsc{text}.

\allsmcp{text} places both uppercase and lowercase letters in small caps. If the font has the c2sc feature, it will be used; if not, the letters will be converted to lowercase first. Then, smcp is applied.

• If you require small caps, but your font does not have them, you can fake them by scaling uppercase. But this is usually unattractive. A better method is described in the fontspec package documentation: You may appoint a different font that will provide small caps.

• If you insist on using old-fashioned TeX fonts, which do not have the smcp Open Type feature, then you may need to write this code in the Preamble:

\let\smcp\oldscshape\relax
\let\textsc\oldscshape\relax
\let\scshape\oldscshape\relax

The above code sends you in a time machine back to the Twentieth Century.

\acronym{text} is special to novel. If the NovelDeco font is installed, uppercase letters in the text will be converted to "mid-size capitals," and lowercase will remain lowercase, but with ascenders slightly reduced. The purpose of this command is so that acronyms such as NASA and NATO blend well with surrounding text. I have found that whereas ordinary uppercase is too conspicuous, small caps also looks wrong. Example:

The NATO alliance...\par
The \acronym{NATO} alliance...\par
The \allsmcp{NATO} alliance...\par

Whether or not to use the \acronym command, is up to you. Since its letters are drawn from the NovelDeco font, rather than your main font, there may be some stylistic differences if your parentfont is not Libertinus Serif.

\squeeze{text} narrows the text by one percent.

\crush{text} narrows the text by two percent.

These commands are for emergency use only. They temporarily apply a horizontal scale to the letters, so that they are very slightly narrower than usual. Inter-word spacing is not necessarily changed (it can even grow larger), because spacing is calculated during paragraph layout. The enclosed text may span paragraphs.

The commands do not compound. That is, placing one crush inside another will not narrow the text by four percent. But any adjustments made by microtype will still apply, and may compound with these commands.

The only application for these commands is when a paragraph does not quite fit where you would like it to be. For example: Perhaps one short word, or part of a word, dangles to an extra line. Try \squeeze first, then \crush, for all or most of the paragraph. These do not always produce a useful effect, since the actual paragraph layout depends on other factors. But if you are lucky, then the paragraph re-flows to a better layout.

• If the commands do not help, then be sure to remove them, or the typography will be worse than it was before!

{\fontcode text} where fontcode is a previously-defined built-in or user-defined font command.

You locally change the font in use, by calling its font command. For example, \subchfont is a built-in command, automatically used in certain macros. You can use it yourself:

{\subchfont text} % Note that the braces go on the outside.

Likewise, if you used \NewFontFamily or \NewFontFace to define your own font, giving it the command name \mynewfont, then you would use it like this:

{\mynewfont text} % Again, the braces go on the outside.

{\addfontfeature{feature} text} for changing Open Type features.

If a font is already defined, you can locally add one or more features. For example, suppose that you defined \mynewfont without using its Open Type feature ss03 (we will assume that the font actually has this feature). In a particular place, you wish to activate ss03. The code looks like this:

{\mynewfont\addfontfeature{RawFeature=+ss03} text} % Note + sign. Braces outside.

That limits the scope of the feature to the included group, in TeX terminology.

You can modify features used by any defined font, whether you defined it yourself or it is one of the novel built-in font commands. You can add more than one feature. If the requested feature does not exist in the font, then your request will be ignored.

What about removing a feature that is already present in the font definition? Suppose you defined \mynewfont so that Numbers=OldStyle is used. On a local basis, can you change the style of numbers? Usually, yes. You could use Numbers=Lining to over-ride the feature. Sometimes it depends on how you chose the features, as a matter of syntax. See the fontspec documentation.

For this effect, the graytext class option must be used, or any color command will be ignored, and its included text will be rendered in black (even if the requested color is white). Images are unaffected.

\color{name} or \color[gray]{scale} is used within a group. All text will be in grayscale, from where the command is placed, until the group ends. If the command is not at the beginning of the group, it does not affect prior text. Example:

\begingroup
This text is unaffected by the following command.
\color{lightgray} % Alternative: \color[gray]{number 0 to 1}
This text is in lightgray, until the group ends.
\endgroup

\textcolor{text} or \textcolor[gray]{scale}{text} affects only the included text.

\textcolor{gray7}{This text is in color gray7, a medium-light gray.}
\textcolor[gray]{0.32}{This text is in 32 percent gray.}

Whether you use \color or \textcolor, there are two ways to specify the color: by name, or by scale, using package xcolor syntax.

• The following color names are accepted, and will be converted to some shade of gray (from dark to light): black, blue, violet, darkgray, purple, red, teal, olive, magenta, gray, brown, green, orange, cyan, lightgray, lime, pink, yellow, white.

The above names are often not helpful, so novel also allows you to use these: black, gray1, gray2, gray3, gray4, gray5, gray6, gray7, gray8, gray9, white. The scale is not uniform (gray5 is not 50 percent gray), but the values are usefully spaced.

Other names "known to dvips," such as Apricot, do not work. An error will result.

• To specify color by scale, follow the color command with [gray]{scale}, where the scale is a number between 0 (black) and 1 (white). Example: \color[gray]{0.63}.

• If you use xcolor syntax to request an rgb color, it will be converted to 1-channel grayscale.

• Absolutely no transparency!

• The \raisebox command does not play well with novel. It can be used, but there is the great risk of disturbing the line grid. Use \charscale instead.

• If you must use the \vspace command (or its relative, \vskip) be sure to measure the space in units of \nbs (normal baseline skip).

• Avoid the minipage and parbox environments. These are likely to disrupt the line grid. When that happens, restoring the grid is your own responsibility. You may find the adjustwidth and linegap environments to be useful, as described in the next section.

• In most cases, do not use the center environment. You may use {\centering ... \par} instead, for a single paragraph. The center environment may cause the text to go off-grid. That's not a problem on some stand-alone pages where you don't care about the grid, but it is ugly in the middle of running text.

• Avoid positioning via the textblock or textblock* environments (package textpos). Actually, novel uses that code internally. But if not handled with more care than usual, positioned text can disrupt the baseline and position of surrounding text. You may use \charscale to position small amounts of text safely, but not on a paragraph basis. Images have built-in positioning arguments, which may also be useful for positioned text.

• Do not use floats. If you try, then novel may intercept them with a nasty message.

• Do not use tables. Nasty message will result. But novel has its own provision for a Table of Contents, and the code may be used for other purposes as well.

• If you use images, then use only the image commands provided by novel: \InlineImage and \BlockImage.

• There is no built-in provision for a two-column environment. However, you can place text alongside an image.

This is not a complete list. It merely addresses a few situations that you are likely to encounter, without the need for elaborate code, and without loading additional packages.

\begin{adjustwidth}{left}{right} ... \end{adjustwidth}

This environment, from the included changepage package, is the most useful way to create block indents. The included text may have multiple paragraphs and span pages. The left and right arguments are lengths, being the amount to increase the respective margins from their normal values.

\begin{linegap}{integer} ... \end{linegap}

This environment creates a gap of integer lines. Then, the following text will be on-grid, regardless of what you place inside the environment. If the contents are small relative to the gap, there will be a lot of white space. If the contents are too large, then they will overflow the gap and overlap the following text, rather than push it down.

This environment is particularly useful when it contains highly styled text, because you don't have to calculate a \vspace to restore the following line grid. Pick a likely gap, then make it larger or smaller as desired.

The environment fails if it crosses a page boundary. There will be no error, merely an undesired result. Be sure to plan your writing so that the linegap remains on a single page.

If you use a non-integer linegap, it will be processed that way, but there will be a warning in the log file. The only reason it is not an error, is that a non-integer linegap might be useful on some display pages.

\rotatebox{text}

This command is from the included graphicx package. It rotates the text, about an axis at the lower left of the text. If you wish to rotate the text about its center, use \rotatebox[origin=c]{text}.

This command plays nicely with \charscale, so you can size, rotate, and displace a string of text (single line). It makes a difference which command goes inside the other. More likely than not, the effect you want is obtained by using \charscale on the outside, with \rotatebox inside.

To avoid disturbing the line grid, \rotatebox is best used inside the linegap environment.

\phantom{text}

Does not print the text, but leaves a horizontal gap as if the text were there.

\mbox{text}

Prevents a specific word from being hyphenated, or to ensure that two consecitive words always appear on the same line. The box will not be broken across lines. However, it may overhang the margin, which is undesirable.

\makebox[width][alignment]{text}

Puts the text in a box of fixed width (usually much larger than the text's own width) and aligns it left, center, or right in the box. If using this across the full textwidth, be sure to precede it with \noindent to prevent offset.

\hspace{length}

Creates a horizontal space.

\strut

Places an invisible box with no width, which has the height and depth of the font. Sometimes used as a placeholder.

\stake

Placeholder like \strut, but does not occupy the height or depth of the font.

\indent, \noindent, \forceindent, \backindent

Commands \indent and \noindent have their ordinary TeX meaning. Sometimes, the action of other macros will remove an indent that should be there, or add an indent that should not be there. If writing \indent and \noindent do not fix the problems, use \forceindent and \backindent as more forceful solutions.

\hfill

Pushes the following text to the right end of the line. Actual effect depends on what precedes and follows. Sometimes a \strut is needed at the beginning of a line.

\vfill, \vfill*

Pushes the following text to the bottom of the page. Actual effect depends on paragraph organization. Use the asterisk if there is nothing printed above this command, on the same page.

\kern

Tweaks the spacing between consecutive letters. A tricky command, as it is always used without braces, and makes the surrounding text illegible in your source code. If you use it, the units of length should always be em, rather than pt. Example:

It wa\kern-.02ems a da\kern.03emrk and stormy ni\kern-.004emght.

The \kern command is not useful for globally changing kern values.

\bigemdash[yoffset,thickness]

Simulates a "big emdash" of variable length. Horizontal width is from command to end of line marked by \par, or limited by a fixed width such as created by \makebox. If placed in other situations, the line may be missing or too long.

The line's normal position is at the typical height of an emdash, which is slightly above half the height of lowercase x. Its normal width is typical of an emdash width. Thus, the line is like an extended emdash.

You can tweak the line using its optional argument. The first value tweaks the vertical position by a scale factor, where 1 equals its normal position (so that 0.9 places it a little lower). The second argument, if used, is separated from the first by a comma. It tweaks the line thickness by a scale factor (so that 0.9 makes it a little thinner).

Since \bigemdash allows an optional argument, but no mandatory argument, you may not immediately follow it with text. If it is boxed, then first close the box; otherwise, end the paragraph.

\straightquote, \straightdblquote

These insert the single quote (apostrophe), and double quote, as ASCII straight characters. Prevents TeX (or your software) from automatically substituting the curly characters.

This command may only be used in the Preamble. It globally sets the kind of marker to be used. Default is real and asterisk.

The optional argument applies real or fake markers, as described above.

The required argument applies only to markers created by the \footnote command. It does not apply to endnotes, or to manually placed superscripts and subscripts.

• asterisk (default) uses symbolic footnote markers. The symbols rotate in order * † ‡ § ** †† ‡‡ §§ until reset.

• dagger uses symbolic footnote markers, but without asterisks. The symbols rotate in order † ‡ § †† ‡‡ §§ until reset.

• number uses numerical footnote markers. Due to possible conflict, the log file will have a "Big Bad Warning" if you use a numerical footnote marker in a document that also has endnotes, since endnotes always have numerical markers.

Within the body of the document, you may use \realmarker and \fakemarker to change your preference. The choice remains in effect until you change it again.

The option real/fake applies wherever superscripts or subscripts are used, not just for footnotes or endnotes.

Footnotes appear at the bottom of the page where they are placed. There is no option to accumulate them to another location. Very long footnotes might straddle pages, but that situation is uncommon in fiction (and not within your control).

\footnote[option]{text}

The \footnote command places a superscript "marker" at the place where the command is written. The included text is written at the bottom of the page, beneath a short line that separates it from the main text. The footnote text is at smaller size and line skip, and has the same marker placed before it (but not as superscript). All of this is automatic.

You may, if necessary, apply some style to the footnote text. For example, if you are referring to another work, its title might be in italics.

Normally, no option is used. If you use option 0 (zero) then no marker will appear in the main text, and the footnote text will also appear without a marker. The count of markers is not incremented. This applies whether the markers are numbers or symbols.

• Trick: If you are using footnote numbers, but wish to place a special footnote using a symbol, you can do it manually. For example:

He blinked.*\footnote[0]{* No, he didn't.}

• Avoid use of underlining, as it is considered bad style, due to visual distraction. If necessary, use italics.

Endnotes are sometimes used in fiction. In general, the endnotes are also fictional.

You may use both footnotes and endnotes, if you wish. However, if endnotes are used, then footnotes must be marked by symbols, rather than by numbers, to avoid confusion. The counts for endnotes and footnotes are independent.

6.5.2.1 \endnote, \ResetEndnote, \memo{}

6.5.2.2 \endnotetext{number}{text}

The novel image-handling commands require the png or jpg formats, exclusively. If you already have a raster image in another format, such as tif, you must convert it. If you have a vector image, such as eps or svg format, you must rasterize it. This must all be done in an external graphics program, such as GIMP or Inkscape. It cannot be done from within TeX.

Both jpg and png images can store their resolution internally. When you export images from a graphics program, be sure that the resolution is stored. This may or may not be an option that you need to check, in an export dialog box.

Be sure to flatten the image before you export it to png. With jpg, the exported image is always flattened.

When using grayscale, be sure to convert the image to only grayscale, which has only one color channel. In GIMP, this is achieved with the Image>Mode menu. Note that an RGB or CMYK image "that only uses grays" is not the same thing, since these have three or four color channels!

When using pure black/white (png only), you can get there with GIMP's Image>Mode>Indexed menu.

Note that within the finished PDF file, images are stored in an internal format which differs from your input. If your image is in the wrong format, or contains transparency, it will not be detected by TeX.

Image resolution is measured in dots per inch (dpi). In a graphics program, it may be called pixels per inch (ppi). And, if the software was written in Europe or Asia, it may be measured in pixels per centimetre!

When you prepare an image in a graphics program, you must first ensure that it is at the correct resolution. Most U.S. print services require 300dpi for grayscale images (or color covers), 600dpi for black/white line art. Then, it is a simple matter of multiplying the image dimensions (in inches) by the resolution, to get the image dimensions in pixels.

Example: A grayscale image that will print at 1.25" wide by 1.5" high, must be 375 pixels wide by 450 pixels high, at 300dpi. If the original image is at different dimensions, it must be cropped and scaled to the correct size. However, it is not enough to finish the image at 375x450 pixels. You must specifically set the resolution to 300dpi. If you fail to set the resolution correctly, then the image will not appear at its correct size in your PDF.

In GIMP, you can use the Image>Scale Image dialog box to ensure that the image is at the proper size and resolution. When you change any of the settings, others may change in response. Be sure that all settings are satisfactory.

When you export your image, you may be asked whether or not to save the resolution. Be sure to save it.

Example dialog boxes for grayscale png exported from GIMP:

Your image may contain "private metadata" that records how the image was created, especially if it was converted from a digital photograph. For example, EXIF and IPTC may contain a lot of private metadata. The image may also contain color profile data that is inappropriate for a grayscale or black/white image.

Many commercial printing standards require that such private metadata be removed from images. The PDF File Data should speak for everything, not the individual images.

The simplest way to remove private metadata is via a command-line program: Use either ImageMagick, or GraphicsMagick. These programs work on Windows, Linux, and (so I am told) OS/X. Specific code is shown in the workflow examples, below. If you ever re-work an image, you must once again remove its private metadata.

To create a suitable line art image in GIMP, with post-processing in ImageMagick or GraphicsMagick, follow these instructions:

(1) Open the image file in GIMP. If it is a vector format (such as svg or pdf), ensure that it is resolved to 600 pixels/in with anti-aliasing, and that its size is the same as your intended dimensions (measured in inches or mm). You can also use a larger size, if you intend to crop some of the periphery; or you can use a smaller size, if you intend to make it larger by adding white space.

(2) If you need to crop to a smaller size, or add to a larger size, do it now. Image>Canvas Size, with Resize Layers:All.

(3) If you need to edit the image, by adding or removing anything, do it now.

(4) Image>Flatten Image.

(5) Image>Mode>Grayscale.

(6) View>Zoom to an appropriate amount that shows the image at approximately its print size. That depends on your monitor's resolution. On my own computer, 25% is about right. Does not need to be exact.

(7) Colors>Threshold. If necessary, adjust the slider until the image looks best. It might be that the default 50% works for you. If you move it to the left, then black areas will become a little narrower. If to the right, black areas will become a little thicker.

(8) Image>Mode>Indexed>(black and white, 1-bit palette). Remove unused colors from colormap. No dithering.

(9) File>Export. Use png (Portable Network Graphics) extension. When the dialog box appears, ensure that "save resolution" is checked. Use the highest level of compression offered. The exported image may look poor in some image viewers, because those viewers do not read the full resolution. Not a problem. The image will look good when seen in better viewers.

(10) Command-Line Terminal, using GraphicsMagick or ImageMagick:

GraphicsMagick, Windows or Linux: gm mogrify -strip nameofimage.png

ImageMagick, Windows: magick mogrify -strip nameofimage.png

ImageMagick, Linux: mogrify -strip nameofimage.png

(11) Sanity check:

GraphicsMagick, Windows or Linux: gm identify -verbose nameofimage.png

ImageMagick, Windows: magick identify -verbose nameofimage.png

ImageMagick, Linux: identify -verbose nameofimage.png

After a few moments, the Terminal will show some information. It should say that the depth is 1-bits per pixel, that the only color is gray (1 bits), that the resolution is 236.22 pixels per centimeter (which is 600dpi). The page geometry, divided by 600, will give you the image size in inches (thus, a page geometry of 900+1200+0+0 represents an image 1.5" wide x 2" high).

To produce a suitable grayscale image in GIMP, with post-processing in ImageMagick or GraphicsMagick, follow these instructions:

(1) Open the image file in GIMP. If it is a vector format (such as svg or pdf), ensure that it is resolved to 300 pixels/in with anti-aliasing, and that its size is the same as your intended dimensions (measured in inches or mm). You can also use a larger size, if you intend to crop some of the periphery; or you can use a smaller size, if you intend to make it larger by adding white space.

(2) If you need to crop to a smaller size, or add to a larger size, do it now. Image>Canvas Size, with Resize Layers:All.

(3) If you need to edit the image, by adding or removing anything, do it now.

(4) Image>Flatten Image.

(5) Image>Mode>Grayscale.

(6) View>Zoom to an appropriate amount that shows the image at approximately its print size. That depends on your monitor's resolution. On my own computer, 50% is about right. Does not need to be exact.

(7) Very dark grays are hard to produce accurately, and so are very light grays. If your image relies on areas that are nearly black and nearly white, then consider adjusting lightness and contrast so that they are not so close to black and white. GIMP's Colors>Levels might be more useful than Colors>Brightness/Contrast.

(8) Ensure that white areas are truly white, not light gray. Use GIMP's Select By Color Tool, with a very low threshold, to choose areas that should be exactly white, then clear them.

(9) File>Export. Use png (Portable Network Graphics) extension. When the dialog box appears, ensure that "save resolution" is checked. Use the highest level of compression offered. The exported image may look poor in some image viewers, because those viewers do not read the full resolution. Not a problem. The image will look good when seen in better viewers.

(10) Command-Line Terminal, using GraphicsMagick or ImageMagick:

GraphicsMagick, Windows or Linux: gm mogrify -strip nameofimage.png

ImageMagick, Windows: magick mogrify -strip nameofimage.png

ImageMagick, Linux: mogrify -strip nameofimage.png

(11) Sanity check:

GraphicsMagick, Windows or Linux: gm identify -verbose nameofimage.png

ImageMagick, Windows: magick identify -verbose nameofimage.png

ImageMagick, Linux: identify -verbose nameofimage.png

After a few moments, the Terminal will show some information. It should say that the type is grayscale, at 8 bits per pixel. The resolution is 118.11 pixels per centimeter (which is 300dpi). The page geometry, divided by 300, will give you the image size in inches (thus, a page geometry of 900+1200+0+0 represents an image 3" wide x 4" high).

Most important of all: Scroll through the information. Near its end, you should see a line like this. The exact wording will depend on your software:

For png images:   png: IHDR.color_type: 0 (grayscale)

For jpg images:   JPEG-Colorspace: 1 and JPEG-Colorspace-Name: GRAYSCALE

The above codes are the only ones that matter. Even if other information claims that the image is grayscale, or lists the colors and they are all shades of gray, that is not good enough. When automated software reviews your PDF for compliance, it will look at the png IHDR.color_type, which must be 0, or the JPEG-Colorspace, which must be 1. Anything else fails. If the information is different, it means that you forget to change the mode from RGB to grayscale before you exported the image.

You may ask: If the image was not really grayscale, can it be converted to grayscale using a simple command line, with ImageMagick or GraphicsMagick? The answer is yes, but don't do it that way. The reason is that there are many ways to map colors (even "gray colors") to true grayscale, and an automated conversion may produce unexpectedly light or dark results. It is better to do the conversion in a graphics program, where you can see what you are doing.

You may scale, rotate, and displace text. This is done using text commands, and the result is still text. When cleverly done, it provides an image-like effect. See section 6 for the possibilities.

This command places the image as an inline element, at 100% scale. The use of png or jpg images, with file extension, is required.

You may offset the image horizontally or vertically. Positive direction is to the right and up. If you use one offset, then you must use both (if necessary, 0em). Any units of length may be used, but the preferred units are em or \textwidth for horizontal, and em or \nbs for vertical.

By default, the top of the image is at the text baseline. If you prefer that the bottom of the image be at the text baseline, use voffset b. Example:

\InlineImage[0em,b]{imageFile.png}

Thanks to the included calc package, offsets may use \widthof{text}, \heightof{text}, or \depthof{text} as lengths. The most useful of these is when you wish to align the top of an image at the X-height of surrounding text:

\InlineImage[0em,\heightof{X}]{imageFile.png}

You may also use \heightof{\Aring} for the tallest accented letter normally found in Latin-1.

Another possibility is to align the bottom of an image with the lowercase descender:

\InlineImage[0em,b-\depthof{gjpqy}]{imageFile.png}

Lengths \imagewidth and \imageheight may also be used for offsets. These lengths remain set, and can be used for other purposes (such as displacing text) until another image is placed.

Offsets may be written as combinations of lengths. If you use the baseline b as above, it must be the first character of voffset. If you use \widthof or \heightof, then its backslash must be the first character unless it is preceded by another length. Thus, an offset of -\heightof{X} is prohibited, but you can use 0pt-\heightof{X}.

Here are some examples of allowable code:

\InlineImage[2.3em, -0.4em]{imageFile.png}
\InlineImage[0pt-\widthof{eee}, b+\heightof{X}-0.2em]{imageFile.jpg}

The image occupies zero vertical space. So, it will not push apart baselines. The next line(s) of text may overlie the image. In some cases, this is desirable. If you do not want that effect, then you can either insert a suitable number of \null lines after the image, or use the \BlockImage command instead.

With the un-starred \InlineImage command, the image occupies its natural horizontal width. That is, text following the image will (if it fits) continue from the right edge of the image. But with the starred \InlineImage* command, the image occupies zero width. In this case, text following the image will continue from the left of the image, and overlie the image.

Having text overlie the image is allowable, and can create some interesting special effects. But when your print service sees what you have done, they may think that it is a mistake. Be sure that you discuss this with them.

If you do allow text to overlie an image, be sure that the image is very light. Repeat: very light. Depending on the printing technology, an image may print darker than you see on your computer screen, and that would make text hard to read. At low print volume, your print service will probably use a technology that is like an industrial-strength home inkjet printer, which can make very sharp text over images. However, if your book goes to volume production, then the technology will be different. Volume printing will use halftoned images, and the text overlay may also become halftoned, which will reduce its sharpness considerably.

Note that an image may be placed in a way that prints some of the image in the margin. That's allowable if your margin is wider than the minimum required for printing clearance. But even if it is not allowable, there will be no complaint or warning from TeX. Use the shademargins class option to ascertain position.

Do not place an image so that it approaches or goes beyond the page edge. That requires a "bleed box" which is not part of novel capabilities. If your book needs that kind of image handling, use a page layout program instead.

Do not place images so that they overlap. Although that is allowed in the PDF, it is likely to be flagged as an error by your print service.

As noted above, an \InlineImage (or its starred version) can be used as the background to any subsequent text. Just as with any other text, whatever sits atop an image must be pure black. If you need text in gray, then you must integrate the rasterized text with the image in an external graphics program, instead of typing the text in your document.

When text overlies an image, it retains its identity as vector paths, as far as the PDF is concerned. In some printing processes, particularly those for very low volume production, the result will be as sharp and dark as ordinary text. However, if your book sells well (or is being prepared to sell well), the image will be converted to halftone, and the text that lies above it may also be converted to halftone. This will reduce the sharpness of the text. It is not necessarily a problem, unless your text is very small or has fine details.

If the text-over-image is part of a title, then the text is naturally large and dark, and will easily withstand halftoning.

Be aware that not all print services will accept text over images. It's not that they cannot do it; after all, your home and office printers can easily do it. The reason they object is because the pages of your book will be handled in small volume by persons who are unfamiliar with the book. Text over an image may be flagged as a printing error or a bug in the PDF. In the advertising world, text over image is routine. But in the world of fictional novels, it is unusual. Don't imagine that they will decide what you meant to do, by reading your book. Also, very sharp image/text, at low printing volume, will not appear as sharp when the technology changes for higher volume production. They know that, but are not sure if you know that.

An inline image occupies zero vertical space. If you place its top near the bottom of the textblock, then the image will happily flow into the lower margin and perhaps off the bottom of the page, without an error (but the print service will reject it). The reason that TeX does not provide an error message is quite simple: Technically, it is not an error! It is just not what you meant to do. The software cannot read your mind. This is where the shademargins class option helps.

Example:

\noindent\InlineImage*[0pt,1.2\nbs]{gavel.png} Lord Withens took a seat at the rear of the courtroom, attempting to be inconspicuous. Yet his effort was in vain, as he was the only person there with a riding crop and spurs.\par
He would have taken the car, had it not been stolen. But that was why he was here, at the arraignment of the person charged with the theft. It was most unfortunate that the accused was his own elderly father; but then, theft ran in the family line.\par

The \BlockImage command places its image as a block element, distinct from the preceding or following text. Be sure to close the preceding paragraph before writing \BlockImage.

The command takes an optional argument with up to three comma-separated values. The first is alignment: l, c, or r for left, center, or right. The second and third values are lengths, hoffset and voffset for the image. Positive offsets are up and right.

Unlike the \InlineImage command, you may not use "b" in the vertical offset. By default top of the image is aligned to the baseline of where text would be, if you did not place an image. You may find that \normalXheight is a useful vertical offset, at it aligns the top of the image to where the X-height would be.

Following the mandatory image filename, an optional argument allows text to appear in the same block as the image. See the next section.

An automatic calculation ensures that text after \BlockImage will comfortably clear the bottom of the image (taking its voffset into account). The text will be on-grid.

Remember that TeX does not care if the image overflows into the margins. Thus, you must review the PDF to ensure that an image does not appear in an unsuitable location. You can visualize margins using the draft,shademargins document class options.

Attention: Changed behavior of \BlockImage in v. 1.46.

The second optional argument of \BlockImage, following the image filename, may contain text that will accompany the image in the same block of space. The text may be more than one paragraph.

If the image is aligned at left or right, then the accompanying text block will be separated from the image by the length of the normal paragraph indent. Within this block, the text will be justified. The usual indentation rules apply.

If the image is aligned at center, then the accompanying text will also be centered (and usually will overlie the image). If you wish to confine the width of this text block to less than the full page text width, then wrap the accompanying text (inside the brackets) in an adjustwidth environment.

Whatever the alignment of image offset, the first line of accompanying text will be vertically located on-grid, where the first line of text would ordinarily sit. If you wish to displace it vertically, then begin the accompanying text with \pushtextdown{length}. It is best to measure the length in terms of the normal baselineskip, \nbs.

Note that you cannot automatically "flow" text around an image in novel class. Although there is a LaTeX package that will do that, parts of it are incompatible with the grid structure of novel, so the package is disallowed. If you cleverly place just the right amount of accompanying text, then you can fill the area next to the image; but this is not automated.

If you write too much accompanying text, it will overflow into the text that follows. There is no automatic increase in the vertical space reserved for the image.

Example:

It was a dark and stormy night. Lord Withens mounted his horse and rode through the darkness, accidentally trampling some petunias in Her Ladyship's garden.\par
\BlockImage[r]{gnome.png}[\pushtextdown{2\nbs}Suddenly he was confronted by an irate gnome, who expressed an opinion in the crudest of language, even for a gnome.\par]
Lord Withens shrugged, and rode on. There was nothing to do but remain silent, this time taking more care. He would speak with his wife about the gnome situation when he returned.\par

• In general, avoid using accompanying text with an image, with the possible exception of titles and other special layouts. There is little need for it in the flowing text of a novel. If you find that you need to use it frequently, then perhaps your book would benefit from a page layout program such as Scribus, instead of TeX.

Create your artwork, including all text (but not barcode or barcode text) using whatever means you have. How you do it is not relevant. To follow these instructions, you must provide the artwork as a flattened (no transparency) PNG format image, at correct size (including bleed area) and resolution (dpi).

Example: Suppose your book is 5.5" x 8.5" softcover, a typical size for American fiction. Based on its page count and paper, you are told that the spine width must be 0.62". The required bleed width, on all four sides, must be 0.125".

Then, the image is 0.125 + 5.5 + 0.62 + 5.5 + 0.125 = 11.87in wide.

The image is 0.125 + 8.5 + 0.125 = 8.75in high.

At 300dpi resolution (a typical value) the image must be 11.87 x 300 = 3561 pixels wide, and 8.75 x 300 = 2625 pixels high. In case of fractional pixels, round upward. The exact pixel count is important.

It is best to use the AdobeRGB1998 color workspace, or equivalent. But you can also use sRGB, or equivalent, and in many cases you will not notice much difference. It is helpful to attach (embed) the workspace color profile in your PNG image, especially if it is something other than AdobeRGB1998 or sRGB. Note that the workspace is NOT the color profile of your monitor (if you have it). If you do not attach (embed) the workspace color profile, then remember whether you were using AdobeRGB1998 or sRGB. If you make a mistake here, it is not fatal. It merely means that some colors (particularly bright greens and blue-greens) will print somewhat brighter or duller than you were expecting.

When done, export it as a flattened PNG.

If you are provided with an image other than PNG, you have to convert it to PNG. No alternative. It is easy enough to use GIMP to convert most image formats. If you have a PDF (which is not already in the required format), GIMP can probably open it as an image. Be sure you import it with the correct resolution. Again, the pixel width and height are important.

Cover Text: These instructions assume that your cover text is rasterized, and is part of the image rather than vector text. In many cases your graphics program has enough text capability to do what you need.

If you wish to use TeX to typeset the cover text, do it this way: Use TeX to create one or more PDFs of the cover text, without images. You can create several blocks of text at different pages sizes, if you wish. Open your background artwork in GIMP. Add a new transparent layer above it. Open a text PDF in GIMP. Choose resolution 300dpi, full antialiasing, grayscale or color. The opened text is now a raster image. Copy it, and paste it atop the transparent layer. Anchor it. Use Color: Color to Alpha: white to alpha filter. That removes the white area from the PDF image. Finally, move the text to where it should go. You can rotate the text 90 degrees, but do not resize it! If you would rather the text be at a different size, then either change it in TeX, or use a different resolution when you open the PDF in GIMP. Note that even when you open the PDF at a different resolution, it will still be at 300dpi when you place it over the 300dpi background image; but its size will change.

Get the 238% total ink limit color profile from Aaron Shepard. Direct link to zipped profile. This does the magic of ensuring that you comply with the 240% ink limit (the profile provides a little headroom, which seems useful). It can also be used for softproofing, if you know how to do that. That's not the only possible profile, but it seems to me that it is the best of several available. A good alternative is from David Blatner. Direct link to icc profile.

Included in the novel documentation extras folder is an open-source color profile for sRGB (file srgb.icc). There is also an open-source Equivalent to AdobeRGB (file crgb.icc). You will be using these when you convert your PNG image to a JPG in CMYK color space. These are re-named, cross-platform files obtained from Elle Stone. I can think of no technical need to use the "real" profiles that may be lurking somewhere in your computer.

Get GraphicsMagick or ImageMagick. On Linux, you can install it via your system package manager. On Windows, get GraphicsMagick Q16 binaries for your system (32 or 64 bit) from SourceForge. I am told that versions for Mac are available via homebrew and macports, but again, I do not have a Mac, so I cannot check that. Or, get ImageMagick Q16 for various platforms from here.

The effect of conversion is explained in Appendix D.

Get ready: Make a copy of the above 238% (or 240%) ink limit color profile, and re-name the copy as 240.icc. I will call your original PNG image yourimage.png. Put this image, the 240.icc file, and files srgb.icc and crgb.icc in a single working folder.

I will also assume that the image resolution is 300dpi. If it is something else, then modify 300 to the other number, when you see it below.

Note: Each of the following boxed commands is a single command line. You may need to scroll the box to the right, or the text may wrap, depending on how you view this file.

Step 1: Open a Terminal (Command Prompt) and navigate to that folder.

In the following commands, the initial "gm" calls GraphicsMagick (same code Linux and Windows). If you prefer ImageMagick, then eliminate the initial "gm" on Linux, or change the "gm" to "magick" on Windows.

The above command may return the message "No color profile available" and generate a zero-byte embedded.icc file. Not a problem. Discard the zero-byte file. But if it generates a non-zero embedded.icc file, keep it.

Step 2: Next (note the file extensions):

Step 3: At this point, there are three possibilities:

(a) You have a non-zero embedded.icc file. In this case:

(b) No embedded.icc, but you know that the artwork was create using AdobeRGB1998 workspace, or equivalent. Then:

(c) Neither of the above. The image is probably sRGB (default). Then:

Step 4: Now, regardless of which possibility you used above:

Step 5: And finally:

Step 6: Now verify your work:

The report should show that the image is CMYK, that it is at 300 pixels/inch, and that there is NO entry for "Profile-color". It will not have anything to say about ink limit, or SWOP.

Up to now, you did not need TeX. Now you do. Be sure that you have a working TeX installation, with the novel document class package.

In folder (your TeX root)/doc/lualatex/novel/extras you will find file "novel-cover-template.tex". Place a copy in the same working folder as your image, and modify it as needed.

\SetMediaSize refers to the full size of the image, with units (usually in). This includes the bleed box. For example, \SetMediaSize{11.87in}{8.75in}.

\SetTrimSize refers to the trimmed image, without bleed. For example, \SetTrimSize{11.62in}{8.5in}.

\SetPDFX[CGATSTR001]{X-1a:2001} will get you "US Web Coated SWOP v2" output intent (most common the USA), and PDF/X-1a:2001 compliance. The Output Intent *.icc color profile will NOT be embedded, because this is almost always the way that print services want it. The PDF contains a code identifying the intent, so there is no need to embed the color profile (and disadvantages to embedding it). In the unlikely event that they demand the Output Intent color profile embedded, put an asterisk on the command.

For compliance you need to use \SetTitle. The title for this file might not be the title of your book, because this file isn't your book. Ask you print service what they want. Same with \SetAuthor.

You can use \SetApplication and \SetProducer to change internal identifiers. Don't use these unless you have a reason.

\SetCoverImage is self-explanatory. Must have file extension jpg, jpeg, JPG, or JPEG. Must be CMYK. Note that TeX cannot verify whether the image is in correct color space. It will compile, even if the image is wrong. Beware!

Do not attempt to write anything between \begin{document} and \end{document}. If you erroneously put some text there, the document will probably compile, and it may even be a valid PDF/X file. However, the resulting text color will almost surely be wrong. If so, the problem will be detected when you submit your file (wrong ink limit or color space).

Compiling may take a little longer than expected. If your TeX system opens a PDF viewer automatically, the image might be missing (or partial) until a few seconds after the viewer opens. Be patient. The file size is large.

Open the PDF in Adobe Reader, which does a trick that no other free software will do, to my knowledge. If your platform does not have Adobe Reader, then find a friend who has it. May be a different platform.

In Adobe Reader, go to Edit, Preferences, Documents. Check "Save As optimizes for fast web view." Then go to Edit, Preferences, Page Display. Check "Show art, trim, and bleed boxes." Then OK.

Open your PDF. You should see a light green box, just inside the image boundary. This shows the Trim Area. This box does not print; it is only for visualization. Go to File, Properties, Fonts. There should be no fonts listed.

Then, "Save As" the same file name (or another one, if you prefer). This will optimize the file. You probably do not need to do that, but it does not hurt, and a few services expect it.

There will always be some difference between color as viewed on your monitor, and color as printed. This is due to the difference between emitted light (your monitor) and reflected light (print). In professional use, the monitor will be calibrated, and operated at a specific screen brightness. The print will be viewed under lighting with a specific color temperature and illumination. Your own colors are not that critical, are they? If you have a calibrated monitor, and view the PDF and images in color-managed software, then what you see on your computer will be reasonably close to what prints. If your monitor is not calibrated, or you are not using color-managed software, then the printed cover may have its greens a bit more colorful than you were expecting. That doesn't mean there will be a greenish cast. It merely means that things such as spring leaves and shamrocks will be a bit greener (in print) than you expected.

Remember the Disclaimer! Your PDF file should now be ready for submission. Only "should," not "is," because this documentation and its accompanying software comes without warranty. Submit at your own risk!

All I can say is that it "worked for me." Screenshots:

A sheet of paper is a single, physical piece of paper, regardless of whether anything is printed on it, or where. Often, paper starts out as a large sheet that will subsequently be cut into smaller sheets. The size doesn't matter. A sheet of paper is, in ordinary understanding, a piece of paper.

In a finished book, a page is one side of a leaf. Thus, there are always exactly twice as many pages are there are leaves. From the manufacturer's point of view, a 240-page (double-sided printing) book has 120 leaves, regardless of whether or not anything is printed on the pages, or how they are numbered.

Imagine a single sheet of US Letter 8.5"Wx11"H paper. It has two pages, since you can print on both sides (even if one is empty).

If you fold that sheet in half across its width, it is still one sheet. But now you have a folio, consisting of two leaves of size 5.5"Wx8.5"H. Each leaf has two pages, so you now have four pages. The fold is at the spine.

Caution: The term "folio" has several different meanings, even in the narrow field of book publishing. It may refer to folded paper, as mentioned above. Or it may refer to a particular sheet size of paper ("foolscap"). Or, it may simply refer to the number printed on a page. This last use of "folio" is discussed in the context of headers and footers.

When a book is commercially manufactured, folios are nested to form signatures. The number of folios in a signature is typically 6 to 16 (that would be 24 to 64 pages). If you were hand-making a single book at home, you might use only one folio per signature.

Now, think about how those pages must be ordered. Suppose a book contains a single signature of 4 folios. The outermost folio is printed with pages 1 and 16 on one side (before folding), and pages 2 and 15 on its other side. The next nested folio has pages 3/14 and 4/13. Nested inside is 5/12 and 6/11. Innermost is 7/10 and 8/9.

Re-arranging the page order is part of the imposition process. The process may involve tweaking the position of each page slightly, to account for offset due to paper thickness when the folios are grouped into signatures. If numerous folios are printed on a larger sheet of paper before cutting, the process may involve rotating folios so that they waste less paper.

Your home printer software can probably do imposition for a single-folio design (four pages). Anything beyond that is left to professional software. You don't need to do it. The print service does it for you, automatically.

Most modern softcover books (and some inexpensive hardcovers) are manufactured using a process named Perfect Binding.

In Perfect Binding, the signatures are clamped together, and the folded area is removed by a cutting machine. Then there are no longer any folios, just individual leaves.

The leaves are glued together at the spine, possibly with reinforcement. Note that the glue is not merely at the edge of the leaves; it penetrates the paper slightly, for better strength. Finally, the cover is attached.

The Perfect Binding method is inexpensive, and is more than adequate for a book that will not receive rough handling. Whether using P.O.D. or a major mass-quantity publisher, most softcover books have been made this way for a long time. Unless you know that your book will be acquired by libraries for addition to their long-term literary collection, there is little reason to have your book manufactured by any other process. The possible exception would be books for small children, which receive rough handling under all circumstances.

Libraries rarely purchase Perfect Binding books (regardless of soft or hard cover) due to lack of durability in a library environment. The exceptions are popular books of no lasting literary value, and how-to books for evolving technologies. These may be obtained by libraries in the cheapest softcover version, because such books will be discarded long before they wear out.

Unless you are sure that your book will be acquired by numerous libraries for long-term addition to their collections, there is little reason to produce a quality hardcover version. Not all P.O.D. services offer hardcover. If you do need a hardcover version, then be sure to carefully inquire regarding its manufacture, and preferably obtain a copy of someone else's book manufactured that way.

The best hardcovers, suitable for library use, are made using the Smyth Sewing process. A Smyth Sewn book has signatures of (typically) 12 or 16 folios. But instead of removing the creases after folding, each signature is sewn through its crease using strong thread, onto a durable spine material. The pages will remain attached, even with hard usage. If the spine material is sufficiently flexible, then the book will lie flat when opened.

Be sure to ask about spine flexibility. The most flexible methods have a soft, curved spine, and are best suited to thick novels. There is also a variety that has a square (flat) spine, which will not quite lie flat when opened. Square-spine Smyth Sewn books seem to be the fashion for documenting intellectual property.

If a major publisher is producing a book that is sure to be ordered by many libraries, due to its author's popularity or maybe a movie tie-in, then it is likely to be released as a Smyth Sewn hardcover. That's because the production method is affordable in quantity. But on a P.O.D. basis, it is a significant expense.

Different bindings must have different ISBNs, because they are different products.

A Perfect Bound softcover cannot be retroactively converted to a Smyth Sewn, Library Bound hardcover. However, it is possible to convert a softcover book to some other kinds of hardcover.

In the simplest method, the soft cover is removed, and a hard cover is glued on. The pages may be shaved in size, so that they don't edge beyond the new cover; or the new cover may be slightly over-sized. Part of the soft cover may be pasted onto the hard cover. That won't look the same as original manufacture, but it avoids creating a dust jacket. Such a hard cover book is no more durable than the softcover, no more flexible at the spine, and perhaps no more resistant to losing pages. But it has a harder cover.

A more advanced method involves adding strength to the glued spine, by sewing it sideways, before adding the hard cover. This results in a book that is less flexible at the spine, but less likely to lose pages. Sometimes, small (page count) books are made this way, when there are a limited number stored for archival purposes, rather than reading. In other words, it does not concern novels.

I only mention this because there are services that will create a hardcover book from an existing softcover copy. That is how they do it. Not very useful to you, except possibly as a personal souvenir. If you want a "real" hardcover book, you want a Smyth Sewn Library Binding with hard cover and curved (flexible) spine. Those must be made directly from PDF, not by re-working an existing book. There are services that will do it (not cheaply), but you must remove the ISBN from wherever it appears, because it does not apply to the book in any binding but its original.

The interior paper may be white or "creme" (off-white). White paper is generally used for nonfiction, and must be used if there are interior color images. Creme paper is generally used for fiction. The idea is that white paper creates an air of authority, whereas creme paper is more comfortable for lengthy reading. I agree. I recently read a novel that was (mis-) printed on white paper, and it was hard to read in a bright area. So, for your novel, choose creme paper.

The cover of a Perfect-Binding softcover book may be produced with a glossy or a matte finish. The fashion is to use glossy for nonfiction, matte for fiction. Again, the idea is authority versus comfort. But you may choose what you wish. "Instant Best-Seller" mass-market paperbacks often have an embossed texture, and maybe even metal foil. In the print-on-demand market, forget it. Don't even think about it.

If you do request a hardcover edition, then it may or may not have a dust jacket (costs extra). The general rule is that books for libraries have a dust jacket, which will be encased in plastic by the library. Books for documentation generally have no dust jacket, with the title and other info embossed in the cover material, for permanance.

8.3.1.1. Display Pages

8.3.1.2. Chapter-Like Sections

The \frontmatter command must be written immediately after \begin{document}. This sets page numbering to lowercase roman, and performs some other actions.

A book's front matter is about the book. It is not part of the story. The only mandatory front matter is the Title Page and its verso, the Copyright Page. On a practical basis, six pages of front matter are the minimum, and this minimum is often used, even by major publishers. See Appendix B for how to construct basic six-page front matter.

8.3.2.1. Half-Title, page i

8.3.2.2. Blank, page ii

8.3.2.3. Title Page, iii

8.3.2.4. Copyright Page, iv

8.3.2.5. Dedication

8.3.2.6. Epigraph

8.3.2.7. Table of Contents

8.3.2.8. Foreword, Preface, Introduction

8.3.2.9. Acknowledgements

8.3.2.10. Author's Note

8.2.1.11. Map

8.3.2.12. Second Half-Title

8.3.2.13. Final Page of Front Matter

Main matter begins with the \mainmatter command. It forces its page to recto, changes the page numbering to Arabic, and sets the count to 1.

Main matter is the story itself. In some cases, this also includes a fictional setup (Introduction or Prologue) that leads to the longer story line. It also includes material in the aftermath of the story, such as Epilogue, Conclusion, and end notes.

8.3.3.1. Half-Title in Main Matter ?

8.3.3.2. Part Separator in Main Matter

8.3.3.3. Introduction in Main Matter

8.3.3.4. Prologue in Main Matter

8.3.3.5. Chapters

8.3.3.6. Epilogue, Conclusion

8.3.3.7. End Notes, Appendix, etc.

In novel, you must always end your document this way:

\cleartoend
\end{document}

The final page of your book is necessarily verso. P.O.D. printers (and many others) require that it be completely blank, because they will add some manufacturing information to that page. So, if the text of your book finishes on a recto page, \cleartoend adds one blank page. But if the text finishes on a verso page, then \cleartoend adds two blank pages, so that the final page is blank verso.

8.3.4.1. Backmatter? Not!

8.3.4.2. About the Author

When you set an existing font, or create a new font command, you can use either the font (family) name, or the specific font file name. However, there may be a big difference in results!

By family name:

When you use the font family name, you configure all of the Regular, Italic, Bold, and BoldItalic variants. The usual TeX commands will be applied:

Preamble:

\NewFontFamily\rway{Raleway}

Document Body:

{\rway I came, \textbf{I saw}, \emph{I conquered}, \textsc{I left}.}

By file name:

However, if you specify a single font file name, then only that font file will be used, regardless of requests for bold or italic. Small caps will be honored, if the specific file has them:

Preamble:

\NewFontFamily\rwayi{Raleway-Regular-Italic.otf}

Document Body:

{\rwayi I came, \textbf{I saw}, \emph{I conquered}, \textsc{I left}.}

File Name Combinations:

What you write is what you get. You can customize a font family by individually selecting its members. Here is the syntax.

Preamble:

\NewFontFamily\frankenfont[% note the % at linebreaks
  ItalicFont=BaskervilleF-Italic.otf,%
  BoldFont=libertinusserif-semibold.otf,%
  BoldItalicFont=texgyrepagella-bolditalic.otf,%
]{Raleway-Medium.otf}

Document Body:

{\frankenfont I came, \textbf{I saw}, \emph{I conquered}, \textsc{I left}.}

Semibold, Light, etc.: The above syntax is especially useful when you wish to specify semibold (if available), rather than bold. It is also a way to select something such as light versions:

\NewFontFamily\rwlight[% note the % at linebreaks
  ItalicFont=Raleway-Light-Italic.otf,%
  BoldFont=Raleway-Medium.otf,%
  BoldItalicFont=Raleway-Medium-Italic.otf,%
]{Raleway-Light.otf}

Note that the font file in braces is the one used as the "regular" font, in this definition.

Although it is not an Open Type feature, Scale=number may be written like a feature. Scale globally adjusts the size of the font, relative to what it would otherwise be. If a local font size command is also applied, then the result is the compounded value.

If the font will be mingled with other fonts in the same place, Scale can tweak a font so that it does not stand out. If Scale does not exceed 1, then the baselineskip will not be affected. Over 1 risks spreading lines to accomodate the extra size; this should not be done for fonts that will appear in the flowing main text. But over 1 is acceptable for display fonts that do not wrap lines and do not mingle with main text.

Example: Orbitron is a special-purpose font with a mechanical or futuristic look. Its characters are much larger and heavier than would be found in book fonts. If I wish to mingle Orbitron with ordinary text, I might load the font like this:

\NewFontFace\orbitron[Scale=0.6]{Orbitron} % Preamble

Then in the document body, I could use it like this:

The message was brief: {\orbitron DIE, EARTHLING.} Captain Withens smiled as he oiled his ray-gun. But ray-guns don't use oil.\par

Scale is more often used for fonts that are pre-defined to have a special meaning and location, such as chapter titles or page headers. There, the font is not mingled with ordinary text, and it is OK to use a larger Scale. Then, it is not necessary to use local font scaling commands each time the font is used.

Even though the fontspec syntax allows you to set Color and Opacity as pseudo-features when you define a font, the novel class blocks these features. If you attempt to set them as features in a font definition, an error will result. The same applies if you attempt to add them, using \addfontfeature. However, you can still make use of grayscale using ordinary, local color commands.

When you use the graytext class option, you may write text in grayscale, using commands from the included xcolor package. You should only do this for text that is in special locations. For example, if you begin each chapter with a number that is very large, you might wish to make that number gray, so that it does not overpower the main text.

To avoid possible conflicts, the color specification should be chosen from a limited list of names. In order from dark to light: black, blue, violet, darkgray, purple, red, teal, olive, magenta, gray, brown, green, orange, cyan, lightgray, lime, pink, yellow, white. They will all be rendered in shades of gray (of course, black and white will be themselves). You may also choose from this more convenient list: black, gray1, gray2, gray3, gray4, gray5, gray6, gray7, gray8, gray9, white.

If graytext is not chosen as class option, then any color (even white) will be rendered as black.

• Placing light text over a dark background may produce unexpected results in print, even if it looks good in PDF. If you need this effect, it is better to build the text into the background, as a single image.

Once in a while, a font has a glyph that you would like to use, but it is not included in any of the font's Open Type feature sets. So, how to you grab that particular glyph? For example: One well-known font has a glyph named emdash.alt, which is not as long as the font's ordinary emdash glyph, and looks better in text. Unfortunately, emdash.alt is not in any of the font's style sets. Without some trickery, only a program such as InDesign can grab that glyph.

Thanks to fontspec, novel has the necessary trickery. Command \CreateFontFeature creates a fake Open Type feature code, and assigns your list of substitutions to that code. Use comma-separated entries if there are multiple substitutions. For the case in question:

\CreateFontFeature{myem}{emdash=emdash.alt}

The code myem was randomly chosen as something that is not already defined among Open Type feature codes. Once specified, the fake feature may be used in any font definition, such as this:

\SetParentFont[RawFeature=+myem]{Adobe Garamond Pro}

Of course, the fake feature has no effect if the font does not have the substitution character.

This technique may be used for more complicated features, but that is beyond the scope of this documentation. See the fontspec manual if you need to do more.