% $Id: paras.tex,v 1.4 2020/01/01 19:16:14 karl Exp $ % This is part of the book TeX for the Impatient. % Copyright (C) 2003-2020 Paul W. Abrahams, Kathryn A. Hargreaves, Karl Berry. % See file fdl.tex for copying conditions. \input macros \chapter {Commands \linebreak for composing \linebreak paragraphs} \chapterdef{paras} This section covers commands that deal with characters, words, lines, and entire paragraphs. For an explanation of the conventions used in this section, see \headcit{Descriptions of the commands}{cmddesc}. \begindescriptions \section {Characters and accents} %========================================================================== \subsection {Letters and ligatures for European alphabets} \begindesc \xrdef{fornlets} \bix^^{ligatures} ^^{special symbols} ^^{European alphabets} % \ctsx AA {Scandinavian letter \AA} \ctsx aa {Scandinavian letter \aa} \ctsx AE {\AE\ ligature} \ctsx ae {\ae\ ligature} \ctsx L {Polish letter \L} \ctsx l {Polish letter \l} \ctsx O {Danish/Norwegian letter \O} \ctsx o {Danish/Norwegian letter \o} \ctsx OE {\OE\ ligature} \ctsx oe {\oe\ ligature} \ctsx ss {German letter \ss} \explain These commands produce various letters and ligatures from European alphabets. They are useful for occasional words and phrases in these languages---but if you need to typeset a large amount of text in a European language, you should probably be using a version of \TeX\ adapted to that language.\footnote{The \TeX\ Users Group (\xref{resources}) can provide you with information about European language versions of \TeX.} You'll need a space after these commands when you use them within a word, so that \TeX\ will treat the following letters as part of the word rather than as part of the command. You needn't be in \minref{math mode} to use these commands. \example {\it les \oe vres de Moli\`ere} | \produces {\it les \oe vres de Moli\`ere} \endexample \eix^^{ligatures} \enddesc %========================================================================== \subsection {Special symbols} \begindesc ^^{special characters} % \easy\ctspecialx # \ctsxrdef{@pound} {pound sign \#} \ctspecialx $ \ctsxrdef{@bucks} {dollar sign \$} \ctspecialx % \ctsxrdef{@percent} {percent sign \%} \ctspecialx & \ctsxrdef{@and} {ampersand \&} \ctspecialx _ \ctsxrdef{@underscore} {underscore \_} \ctsx lq {left quote \lq} \ctsx rq {right quote \rq} \aux\ctsx lbrack left bracket [ \aux\ctsx rbrack right bracket ] \ctsx dag {dagger symbol \dag} \ctsx ddag {double dagger symbol \ddag} \ctsx copyright {copyright symbol \copyright} \ctsx P {paragraph symbol \P} \ctsx S {section symbol \S} \explain These commands produce various special characters and marks. The first five commands are necessary because \TeX\ by default attaches special meanings to the characters (|#|, |$|, |%|, |&|, |_|). You needn't be in \minref{math mode} to use these commands. You can use the dollar sign in the Computer Modern italic fonts to get the ^{pound sterling} symbol, as shown in the example below. \example \dag It'll only cost you \$9.98 over here, but in England it's {\it \$}24.98. | \produces \dag It'll only cost you \$9.98 over here, but in England it's {\it \$}24.98. \endexample \enddesc \begindesc \cts TeX {} \explain This command produces the \TeX\ logo. Remember to follow it by |\!vs| or to enclose it in a \minref{group} when you want a space after it. \example A book about \TeX\ is in your hands. | \produces A book about \TeX\ is in your hands. \endexample \enddesc \begindesc \cts dots {} \explain ^^{dots} This command produces an ^{ellipsis}, i.e., three dots, in ordinary text. It's intended for use in mathematical writing; for an ellipsis between ordinary words, you should use |$\ldots$| \ctsref{\ldots} instead. Since |\dots| includes its own space, you shouldn't follow it by |\!vs|. \example The sequence $x_1$, $x_2$, \dots, $x_\infty$ does not terminate. | \produces The sequence $x_1$, $x_2$, \dots, $x_\infty$ does not terminate. \endexample \enddesc \see ``Miscellaneous ordinary math symbols'' (\xref{specsyms}). %========================================================================== \subsection {Arbitrary characters} \begindesc \bix^^{characters} \cts char {\} \explain This command produces the character located at position \ of the current font. \example {\char65} {\char `A} {\char `\A} | \produces {\char65} {\char `A} {\char `\A} \endexample \enddesc \begindesc \cts mathchar {\} \explain This command produces the math character whose class, family, and font position are given by \. It is only legal in math mode. \example \def\digger{\mathchar "027F} % Like \spadesuit in plain TeX. % Class 0, family 2, font position "7F. $\digger$ | \produces \def\digger{\mathchar "027F} % class 0, family 2, font position "7F $\digger$ \endexample \enddesc \see |\delimiter| (\xref\delimiter). \eix^^{characters} %========================================================================== \subsection {Accents} \begindesc ^^{accents} \xrdef{accents} % \ctspecialx ' \ctsxrdef{@prime} {^{acute accent} as in \'e} \ctspecialx . \ctsxrdef{@dot} {^{dot accent} as in \.n} \ctspecialx = \ctsxrdef{@equal} {^{macron accent} as in \=r} \ctspecialx ^ \ctsxrdef{@hat} {^{circumflex accent} as in \^o} \ctspecialx ` \ctsxrdef{@lquote} {^{grave accent} as in \`e} \ctspecialx " \ctsxrdef{@quote} {^{umlaut accent} as in \"o} \ctspecialx ~ \ctsxrdef{@not} {^{tilde accent} as in \~a} \ctsx c {^{cedilla accent} as in \c c} \ctsx d {^{underdot accent} as in \d r} \ctsx H {^{Hungarian umlaut accent} as in \H o} \ctsx t {^{tie-after accent} as in \t uu} \ctsx u {^{breve accent} as in \u r} \ctsx v {^{check accent} as in \v o} \explain These commands produce accent marks in ordinary text. You'll usually need to leave a space after the ones denoted by a single letter (see ``Spaces'', \xref{spaces}). \example Add a soup\c con of \'elan to my pin\~a colada. | \produces Add a soup\c con of \'elan to my pin\~a colada. \endexample \margin{`see also' moved to end of group, replacing the one there.} \enddesc \begindesc \cts i {} \cts j {} \explain These commands produce dotless versions of the letters `i' and `j'. You should use them instead of the ordinary `i' and `j' when you are putting an accent above those letters in ordinary text. ^^{dotless letters} Use the ^|\imath| and ^|\jmath| commands (\xref\imath) for dotless `i's and `j's in math formulas. \example long `i' as in l\=\i fe \quad \v\j | \produces long `i' as in l\=\i fe \quad \v\j \endexample \enddesc \begindesc \cts accent {\} \explain ^^{accents} This command puts an accent over the character following this command. The accent is the character at position \ in the current font. \TeX\ assumes that the accent has been designed to fit over a character $1$\thinspace ex high in the same font as the accent. If the character to be accented is taller or shorter, \TeX\ adjusts the position accordingly. You can change \minref{font}s between the accent and the next character, thus drawing the accent character and the character to be accented from different fonts. If the accent character isn't really intended to be an accent, \TeX\ won't complain; it will just typeset something ridiculous. \example l'H\accent94 otel des Invalides % Position 94 of font cmr10 has a circumflex accent. | \produces l'H\accent94 otel des Invalides % Position 94 of font cmr10 has a circumflex accent. \endexample \see Math accents (\xref{mathaccent}). \enddesc %========================================================================== \subsection {Defeating boundary ligatures} \begindesc \bix^^{ligatures} \cts noboundary {} \explain You can defeat a ligature or kern that \TeX\ applies to the first or last character of a word by putting |\noboundary| just before or just after the word. Certain fonts intended for languages other than English contain a special boundary character that \TeX\ puts at the beginning and end of each word. The boundary character occupies no space and is invisible when printed. It enables \TeX\ to provide different typographical treatment to characters at the beginning or end of a word, since the boundary character can be part of a sequence of characters to be kerned or replaced by a ligature. (None of the standard \TeX\ fonts contain this boundary character.) The effect of |\noboundary| is to delete the boundary character if it's there, thus preventing \TeX\ from recognizing the ligature or kern. \eix^^{ligatures} \enddesc %========================================================================== \section {Selecting fonts} \xrdef{selfont} %========================================================================== \subsection {Particular fonts} \begindesc ^^{fonts} % \ctsx fivebf {use $5$-point bold font} \ctsx fivei {use $5$-point math italic font} \ctsx fiverm {use $5$-point roman font} \ctsx fivesy {use $5$-point math symbol font} \ctsx sevenbf {use $7$-point bold font} \ctsx seveni {use $7$-point math italic font} \ctsx sevenrm {use $7$-point roman font} \ctsx sevensy {use $7$-point math symbol font} \ctsx tenbf {use $10$-point bold text font} \ctsx tenex {use $10$-point math extension font} \ctsx teni {use $10$-point math italic font} \ctsx tenrm {use $10$-point roman text font} \ctsx tensl {use $10$-point slanted roman font} \ctsx tensy {use $10$-point math symbol font} \ctsx tenit {use $10$-point italic font} \ctsx tentt {use $10$-point typewriter font} \explain These commands cause \TeX\ to typeset the following text in the specified font. Normally you would enclose one of these font-selecting commands in a group, together with the text to be set in the selected font. Outside of a group a font-selecting command is effective until the end of the document (unless you override it with another such command). \example See how I've reduced my weight---from 120 lbs.\ to {\sevenrm 140 lbs}. | \produces See how I've reduced my weight---from 120 lbs.\ to {\sevenrm 140 lbs}. \endexample \enddesc \begindesc \cts nullfont {} \explain This command selects a font, built into \TeX, that has no characters in it. \TeX\ uses it as a replacement for an undefined font in a family of math fonts. \enddesc %========================================================================== \subsection {Type styles} \xrdef{seltype} \begindesc ^^{type styles} \easy\ctsx bf {use boldface type} \ctsx it {use italic type} \ctsx rm {use roman type} \ctsx sl {use slanted type} \ctsx tt {use typewriter type} \explain These commands select a type style without changing the typeface or the point size.\footnote{ \TeX\ does not provide predefined commands for changing just the point size, e.g., |\eightpoint|. Supporting such commands would require a great number of fonts, most of which would never be used. Such commands were, however, used in typesetting \texbook.} Normally you would enclose one of these type style commands in a group, together with the text to be set in the selected font. Outside of a group a type style command is effective until the end of the document (unless you override it with another such command). \example The Dormouse was {\it not} amused. | \produces The Dormouse was {\it not} amused. \endexample \enddesc \see ``Fonts in math formulas'' (\xref{mathfonts}). %========================================================================== \section {Uppercase and lowercase} \begindesc \bix^^{case conversion} \bix^^{uppercase//conversion to} \bix^^{lowercase//conversion to} \cts lccode {\ \tblentry{number}} \cts uccode {\ \tblentry{number}} \explain The |\lccode| and |\uccode| values for the $256$ possible input characters specify the correspondence between the lowercase and uppercase forms of letters. These values are used by the |\lowercase| and |\uppercase| commands respectively and by \TeX's hyphenation algorithm. \TeX\ initializes the values of |\lccode| and |\uccode| as follows: \ulist\compact \li The |\lccode| of a lowercase letter is the {\ascii} code for that letter. \li The |\lccode| of an uppercase letter is the {\ascii} code for the corresponding lowercase letter. \li The |\uccode| of an uppercase letter is the {\ascii} code for that letter. \li The |\uccode| of a lowercase letter is the {\ascii} code for the corresponding uppercase letter. \li The |\lccode| and |\uccode| of a nonletter are both zero. \endulist Most of the time there's no reason to change these values, but you might want to change them if you're using a language that has more letters than English. \example \char\uccode`s \char\lccode`a \char\lccode`M | \produces \char\uccode`s \char\lccode`a \char\lccode`M \endexample \enddesc \begindesc \cts lowercase {\rqbraces{\}} \cts uppercase {\rqbraces{\}} \explain ^^{case conversion} These commands convert the letters in \, i.e., those tokens with category code $11$, to their lowercase and uppercase forms. The conversion of a letter is defined by its |\lccode| (for lowercase) or |\uccode| (for uppercase) table value. Tokens in the list that are not letters are not affected---even if the tokens are \minref{macro} calls or other commands that expand into letters. \example \def\x{Cd} \lowercase{Ab\x} \uppercase{Ab\x} | \produces \def\x{Cd} \lowercase{Ab\x} \uppercase{Ab\x} \eix^^{case conversion} \eix^^{uppercase//conversion to} \eix^^{lowercase//conversion to} \endexample \enddesc %========================================================================== \section {Interword spacing} \begindesc \bix^^{spaces//interword} \easy\ctsbasic {\\\vs}{} \blankidxref\ctsxrdef{@space} \explain This command explicitly produces an interword space called a ``^{control space}''. A control space is useful when a letter occurs immediately after a control sequence, or in any other circumstance where you don't want two tokens to be run together in the output. The amount of space produced by |\!vs| is independent of preceding punctuation, i.e., its space factor (\xref\spacefactor) is $1000$. Incidentally, if you want to print the `\vs' ^^{visible space} character that we've used here to denote a space, you can get it by typing |{\tt \char `\ }|. \example The Dormouse was a \TeX\ expert, but he never let on. | \produces The Dormouse was a \TeX\ expert, but he never let on. \endexample \enddesc \begindesc \cts space {} \explain This command is equivalent to an input space character. It differs from ^|\ | in that its width \emph{can} be affected by preceding punctuation. \example Yes.\space No.\space Maybe.\par Yes.\!vs!.No.\!vs!.Maybe. | \produces Yes.\space No.\space Maybe.\par Yes.\ No.\ Maybe. \endexample \enddesc \begindesc \ctsact ^^M \xrdef{@newline} \explain This construct produces the ^{end of line} character. It normally has two effects when \TeX\ encounters it in your input: \olist \li It acts as a command, producing either an input space (if it comes at the end of a nonblank line) or a |\par| token (if it comes at the end of a blank line). ^^|\par//from empty line| \li It ends the input line, causing \TeX\ to ignore the remaining characters on the line. \endolist \noindent However, |^^M| does \emph{not} end the line when it appears in the context |`\^^M|, denoting the ASCII code for control-M (the number $13$). You can change the meaning of |^^M| by giving it a different \minref{category code}. See \xrefpg{twocarets} for a more general explanation of the |^^| notation. \example Hello.^^MGoodbye. Goodbye again.\par The \char `\^^M\ character.\par % The fl ligature is at position 13 of font cmr10 \number `\^^M\ is the end of line code.\par Again, \number `^^M is the end of line code, isn't it? % 32 is the ASCII code for a space | \produces {\catcode `\^ = 7 % disable indexing use within this display Hello.^^MGoodbye Goodbye again.\par The \char `\^^M\ character.\par \number `\^^M\ is the end of line code.\par Again, \number `^^M is the end of line code, isn't it?} \endexample \enddesc \begindesc \easy\ctsact ~ \xrdef{@not} \explain The \minref{active character} `|~|', called a ``^{tie}'', produces a normal interword space between two words and links those words so that a line break will not occur between them. You should use a tie in any context where a line break would be confusing, e.g., before a middle initial, after an abbreviation such as ``Dr.'', or after ``Fig.'' in ``Fig.~8''. \example P.D.Q.~Bach (1807--1742), the youngest and most imitative son of Johann~S. Bach, composed the {\sl Concerto for Horn and Hardart}. | \produces \margin{The inversion of dates is deliberate---cf. Peter Schickele.} P.D.Q.~Bach (1807--1742), the youngest and most imitative son of Johann~S. Bach, composed the {\sl Concerto for Horn and Hardart}. \endexample\enddesc \begindesc \easy\ctspecial / \ctsxrdef{@slash} \explain Every character in a \TeX\ \minref{font} has an ``^{italic correction}'' associated with it, although the italic correction is normally zero for a character in an unslanted (upright) font. The italic correction specifies the extra space that's needed when you're switching from a slanted font (not necessarily an italic font) to an unslanted font. The extra space is needed because a slanted character projects into the space that follows it, making the space look too small when the next character is unslanted. The metrics file for a font includes the italic correction of each character in the font. ^^{metrics file//italic correction in} The |\/| command produces an ^{italic correction} for the preceding character. You should insert an italic correction when you're switching from a slanted font to an unslanted font, except when the next character is a period or comma. \example However, {\it somebody} ate {\it something}: that's clear. However, {\it somebody\/} ate {\it something\/}: that's clear. | \produces However, {\it somebody} ate {\it something}: that's clear. However, {\it somebody\/} ate {\it something\/}: that's clear. \endexample \enddesc \begindesc \cts frenchspacing {} \cts nonfrenchspacing {} \explain ^^{interword spacing} \TeX\ normally adjusts the spacing between words to account for punctuation marks. For example, it inserts extra space at the end of a sentence and adds some stretch to the \minref{glue} following any punctuation mark there. The |\frenchspacing| command tells \TeX\ to make the interword spacing independent of punctuation, while the |\nonfrenchspacing| command tells \TeX\ to use its normal spacing rules. If you don't specify |\frenchspacing|, you'll get \TeX's normal spacing. See \xrefpg{periodspacing} for advice on how to control \TeX's treatment of punctuation at the end of sentences. \example {\frenchspacing An example: two sentences. Right? No.\par} {An example: two sentences. Right? No. \par}% | \produces {\frenchspacing An example: two sentences. Right? No.\par} {An example: two sentences. Right? No. \par}% \endexample \enddesc \begindesc \cts obeyspaces {} \explain \TeX\ normally condenses a sequence of several spaces to a single space. |\obeyspaces| instructs \TeX\ to produce a space in the output for each space in the input. |\obeyspaces| does not cause spaces at the beginning of a line to show up, however; for that we recommend the |\obey!-white!-space| command defined in |eplain.tex| (\xref{ewhitesp}). ^^|\obeywhitespace| |\obeyspaces| is often useful when you're typesetting something, computer input for example, in a monospaced font (one in which each character takes up the same amount of space) and you want to show exactly what each line of input looks like. You can use the |\obeylines| command (\xref{\obeylines}) to get \TeX\ to follow the line boundaries of your input. |\obeylines| is often used in combination with |\obeyspaces|. \example These spaces are closed up {\obeyspaces but these are not }. | \produces These spaces are closed up {\obeyspaces but these are not }. \endexample \enddesc \begindesc \cts spacefactor {\param{number}} \cts spaceskip {\param{glue}} \cts xspaceskip {\param{glue}} \cts sfcode {\ \tblentry{number}} \explain These primitive \minref{parameter}s affect how much space \TeX\ puts between two adjacent words, i.e., the ^{interword spacing}. The normal interword spacing is supplied by the current font. As \TeX\ is processing a \minref{horizontal list}, it keeps track of the ^{space factor} $f$ in |\spacefactor|. As it processes each input character $c$, it updates $f$ according to the value of $f_c$, the space factor code of $c$ (see below). For most characters, $f_c$ is $1000$ and \TeX\ sets $f$ to $1000$. (The initial value of $f$ is also $1000$.) When \TeX\ sees an interword space, it adjusts the size of that space by multiplying the stretch and shrink of that space by $f/1000$ and $1000/f$ respectively. Thus: \olist\compact \li If $f=1000$, the interword space keeps its normal value. \li If $f<1000$, the interword space gets less \minref{stretch} and more \minref{shrink}. \li If $f>1000$, the interword space gets more \minref{stretch} and less \minref{shrink}. \endolist % > changed to \ge on the next line after second edition was typeset. % Correction made by A-W production. In addition, if $f\ge2000$ the interword space is further increased by the ``extra space'' parameter associated with the current font. Each input character $c$ has an entry in the |\sfcode| (space factor code) table. The |\sfcode| table entry is independent of the font. Usually \TeX\ just sets $f$ to $f_c$ after it processes $c$. However: \ulist \li If $f_c$ is zero, \TeX\ leaves $f$ unchanged. Thus a character such as `|)|' in \plainTeX, for which $f_c$ is zero, is essentially transparent to the interword space calculation. \li If $f<1000 to \ge here, too, as above. (see the rule above for the case $f\ge2000$). Noncharacter items in a horizontal list, e.g., vertical rules, generally act like characters with a space factor of $1000$. You can change the space factor explicitly by assigning a different numerical value to |\spacefactor|. You can also override the normal interword spacing by assigning a different numerical value to |\xspaceskip| or to |\spaceskip|: \ulist \li |\xspaceskip| specifies the glue to be used when $f\ge2000$; in the case where |\xspaceskip| is zero, the normal rules apply. \li |\spaceskip| specifies the glue to be used when $f<2000$ or when \hbox{|\xspaceskip|} is zero; if |\spaceskip| is zero, the normal rules apply. The stretch and shrink of the |\spaceskip| glue, like that of the ordinary interword glue, is modified according to the value of $f$. \endulist See \knuth{page~76} for the precise rules that \TeX\ uses in calculating interword \minref{glue}, and \knuth{pages~285--287} for the adjustments made to |\spacefactor| after various items in a horizontal list. \eix^^{spaces//interword} \enddesc %========================================================================== \section {Centering and justifying lines} \begindesc \bix^^{centering} \bix^^{flush left} \bix^^{flush right} \bix^^{justification} \easy\cts centerline {\} \cts leftline {\} \cts rightline {\} \explain The |\centerline| command produces an \minref{hbox} exactly as wide as the current line and places \ at the center of the box. The |\leftline| and |\rightline| commands are analogous; they place \ at the left end or at the right end of the box. If you want to apply one of these commands to several consecutive lines, you must apply it to each one individually. See \xrefpg{eplaincenter} for an alternate approach. Don't use these commands within a paragraph---if you do, \TeX\ probably won't be able to break the paragraph into lines and will complain about an overfull hbox. \example \centerline{Grand Central Station} \leftline{left of Karl Marx} \rightline{right of Genghis Khan} | \produces \centerline{Grand Central Station} \leftline{left of Karl Marx} \rightline{right of Genghis Khan} \eix^^{centering} \eix^^{flush left} \eix^^{flush right} \eix^^{justification} \endexample \enddesc \begindesc \easy\cts line {\} \explain This command produces an \minref{hbox} containing \. The hbox is exactly as wide as the current line, i.e., it extends from the right margin to the left margin. \example \line{ugly \hfil suburban \hfil sprawl} % Without \hfil you'd get an `underfull box' from this. | \produces \line{ugly \hfil suburban \hfil sprawl}% \endexample \enddesc \begindesc ^^{overlapping text} \cts llap {\} \cts rlap {\} \explain These commands enable you to produce text that overlaps whatever happens to be to the left or to the right of the current position. |\llap| backspaces by the width of \ and then typesets \. |\rlap| is similar, except that it typesets \ first and then backspaces. |\llap| and |\rlap| are useful for placing text outside of the current margins. Both |\llap| and |\rlap| do their work by creating a \minref{box} of zero~width. You can also use |\llap| or |\rlap| to construct special characters by ^{overprinting}, but don't try it unless you're sure that the characters you're using have the same width (which is the case for a monospaced font such as ^|cmtt10|, the Computer Modern $10$-point ^{typewriter font}). ^^{Computer Modern fonts} \example \noindent\llap{off left }\line{\vrule $\Leftarrow$ left margin of examples\hfil right margin of examples $\Rightarrow$\vrule}\rlap{ off right} | \produces \noindent\llap{off left }\line{\vrule $\Leftarrow$ left margin of examples\hfil right margin of examples $\Rightarrow$\vrule}\rlap{ off right} \endexample %\example %{\tt O\llap{!|}} %| %\produces %{\cm \tt O\llap{\char `|}} %\endexample \nobreak % don't lose the \see \enddesc \see |\hsize| (\xref{\hsize}). %========================================================================== \section {Shaping paragraphs} \subsection {Starting, ending, and indenting paragraphs} \begindesc \bix^^{paragraphs//shaping} \ctspecial par \ctsxrdef{@par} \explain This command ends a paragraph and puts \TeX\ into \minref{vertical mode}, ready to add more items to the page. Since \TeX\ converts a blank line in your input file into a |\par| \minref{token}, you don't ordinarily need to type an explicit |\par| in order to end a paragraph. An important point is that |\par| doesn't tell \TeX\ to start a paragraph; it only tells \TeX\ to end a paragraph. \TeX\ starts a paragraph when it is in ordinary vertical mode (which it is after a |\par|) and encounters an inherently horizontal item such as a letter. As part of its ceremony for starting a paragraph, \TeX\ ^^{paragraphs//starting} inserts an amount of vertical space given by the parameter |\parskip| (\xref{\parskip}) and indents the paragraph by a horizontal space given by |\parindent| (\xref{\parindent}). You can usually cancel any interparagraph space produced by a |\par| by giving the command |\vskip -\lastskip|. It can often be helpful to do this when you're writing a \minref{macro} that is supposed to work the same way whether or not it is preceded by a blank line. You can get \TeX\ to take some special action at the start of each paragraph by placing the instructions in ^|\everypar| (\xref\everypar). See \knuth{pages~283 and 286} for the precise effect of |\par|. \example \parindent = 2em ``Can you row?'' the Sheep asked, handing Alice a pair of knitting-needles as she was speaking.\par ``Yes, a little% ---but not on land---and not with needles---'' Alice was starting to say, when suddenly the needles turned into oars. | \produces \parindent = 2em ``Can you row?'' the Sheep asked, handing Alice a pair of knitting-needles as she was speaking.\par ``Yes, a little% ---but not on land---and not with needles---'' Alice was starting to say, when suddenly the needles turned into oars. \endexample \enddesc \begindesc \cts endgraf {} \explain This command is a synonym for the ^|\par| primitive command. It is useful when you've redefined ^|\par| but still want access to the original definition of |\par|. \enddesc \begindesc \cts parfillskip {\param{glue}} \explain ^^{paragraphs//glue at end of} This parameter specifies the horizontal glue that \TeX\ inserts at the end of a paragraph. The default value of |\parfillskip| is |0pt plus 1fil|, which causes the last line of a paragraph to be filled out with blank space. A value of |0pt| forces \TeX\ to end the last line of a paragraph at the right margin. \enddesc \bix^^{indentation} \begindesc \easy\cts indent {} \explain If \TeX\ is in vertical mode, as it is after ending a paragraph, this command inserts the ^|\parskip| interparagraph glue, puts \TeX\ into horizontal mode, starts a paragraph, and indents that paragraph by |\parindent|. If \TeX\ is already in horizontal mode, this command merely produces a blank space of width |\parindent|. Two |\indent|s in a row produce two indentations. ^^{indentation} As the example below shows, an |\indent| at a point where \TeX\ would start a paragraph anyway is redundant. When \TeX\ is in vertical mode and sees a letter or some other inherently horizontal command, it starts a paragraph by switching to horizontal mode, doing an |\indent|, and processing the horizontal command. \example \parindent = 2em This is the first in a series of three paragraphs that show how you can control indentation. Note that it has the same indentation as the next paragraph.\par \indent This is the second in a series of three paragraphs. It has \indent an embedded indentation.\par \indent\indent This doubly indented paragraph is the third in the series. | \produces \parindent = 2em This is the first in a series of three paragraphs that show how you can control indentation. Note that it has the same indentation as the next paragraph.\par \indent This is the second in a series of three paragraphs. It has \indent an embedded indentation.\par \indent\indent This doubly indented paragraph is the third in the series. \endexample \enddesc \begindesc \easy\cts noindent {} \explain If \TeX\ is in vertical mode, as it is after ending a paragraph, this command inserts the ^|\parskip| interparagraph glue, puts \TeX\ into horizontal mode, and starts an unindented paragraph. It has no effect in horizontal mode, i.e., within a paragraph. Starting a paragraph with |\noindent| thus cancels the indentation by |\parindent| that would normally occur there. ^^{indentation} A common use of |\noindent| is to cancel the indentation of the first line of a paragraph when the paragraph follows some displayed material. \example \parindent = 1em Tied round the neck of the bottle was a label with the words \smallskip \centerline{EAT ME}\smallskip \noindent beautifully printed on it in large letters. | \produces \parindent = 1em Tied round the neck of the bottle was a label with the words \smallskip \centerline{EAT ME}\smallskip \noindent beautifully printed on it in large letters. \endexample \enddesc \margin{{\tt\\textindent} moved here from later in the section.} \begindesc \cts textindent {\} \explain ^^{indentation} This command tells \TeX\ to start a paragraph and indent it by |\par!-indent|, as usual. \TeX\ then right-justifies \ within the indentation and follows it with an en space (half an em). \PlainTeX\ uses this command to typeset footnotes (\xref\footnote) ^^{footnotes//using \b\tt\\textindent\e\ with} and items in lists (see |\item|, \xref\item). \example \parindent = 20pt \textindent{\raise 1pt\hbox{$\bullet$}}% You are allowed to use bullets in \TeX\ even if you don't join the militia, and many peace-loving typographers do so. | \produces \parindent = 20pt \textindent{\raise 1pt\hbox{$\bullet$}}% You are allowed to use bullets in \TeX\ even if you don't join the militia, and many peace-loving typographers do so. \endexample\enddesc \begindesc \cts parindent {\param{dimen}} \explain This \minref{parameter} specifies the amount by which the first line of each paragraph is to be indented. ^^{indentation} As the example below shows, it's a bad idea to set both |\parindent| and ^|\parskip| to zero since then the paragraph breaks are no longer apparent. \example \parindent = 2em This paragraph is indented by 2 ems. \par \parindent=0pt This paragraph is not indented at all. \par Since we haven't reset the paragraph indentation, this paragraph isn't indented either. | \produces \parindent = 2em This paragraph is indented by 2 ems. \par \parindent=0pt This paragraph is not indented at all. \par Since we haven't reset the paragraph indentation, this paragraph isn't indented either. \endexample\enddesc \begindesc \cts everypar {\param{token list}} \explain \TeX\ performs the commands in \ whenever it enters horizontal mode, e.g., when it starts a paragraph. By default |\everypar| is empty, but you can take extra actions at the start of every paragraph by putting the commands for those actions into a token list and assigning that token list to |\everypar|. \example \everypar = {$\Longrightarrow$\enspace} Now pay attention!!\par I said, ``Pay attention!!''.\par I'll say it again!! Pay attention!! | \produces \everypar = {$\Longrightarrow$\enspace} Now pay attention!\par I said, ``Pay attention!''.\par I'll say it again! Pay attention! \endexample \enddesc \secondprinting{\vfill\eject} %========================================================================== \subsection {Shaping entire paragraphs} \begindesc \margin{This command was also described in the `Pages' chapter. The description here now combines the two earlier descriptions.} \bix^^{line breaks//and paragraph shape} \easy\cts hsize {\param{dimen}} \explain This \minref{parameter} specifies the current ^{line length}, i.e., the usual width of lines in a paragraph starting at the left margin. A great many \TeX\ commands, e.g., |\centerline| (\xref{\centerline}) and |\hrule| (\xref{\hrule}), implicitly use the value of |\hsize|. By changing |\hsize| within a group you can change the width of the constructs produced by such commands. If you set |\hsize| within a \minref{vbox} that contains text, the vbox will have whatever width you've given to |\hsize|. ^^{vboxes//width determined by \b\tt\\hsize\e} \PlainTeX\ sets |\hsize| to |6.5in|. \example {\hsize = 3.5in % Set this paragraph 3.5 inches wide. The hedgehog was engaged in a fight with another hedgehog, which seemed to Alice an excellent opportunity for croqueting one of them with the other.\par}% | \produces {\hsize = 3.5in The hedgehog was engaged in a fight with another hedgehog, which seemed to Alice an excellent opportunity for croqueting one of them with the other.\par}% \doruler{\8\8\8\tick\1\tick\2\tick\1\tick\3}{3.5}{in} \nextexample \leftline{\raggedright\vtop{\hsize = 1.5in Here is some text that we put into a paragraph that is an inch and a half wide.}\qquad \vtop{\hsize = 1.5in Here is some more text that we put into another paragraph that is an inch and a half wide.}} | \produces \leftline{\raggedright\vtop{\hsize = 1.5in Here is some text that we put into a paragraph that is an inch and a half wide.}\qquad \vtop{\hsize = 1.5in Here is some more text that we put into another paragraph that is an inch and a half wide.}} \endexample \enddesc \begindesc \easy\cts narrower {} \explain ^^{paragraphs//narrow} This command makes paragraphs narrower, increasing the left and right margins by |\parindent|, the current paragraph ^{indentation}. It achieves this by increasing both |\leftskip| and |\rightskip| by |\parindent|. Normally you place |\narrower| at the beginning of a \minref{group} containing the paragraphs that you want to make narrower. If you forget to enclose |\narrower| within a group, you'll find that all the rest of your document will have narrow paragraphs. |\narrower| affects just those paragraphs that end after you invoke it. If you end a |\narrower| group before you've ended a paragraph, \TeX\ won't make that paragraph narrower. \example {\parindent = 12pt \narrower\narrower\narrower This is a short paragraph. Its margins are indented three times as much as they would be had we used just one ``narrower'' command.\par} | \produces {\parindent = 12pt \narrower\narrower\narrower This is a short paragraph. Its margins are indented three times as much as they would be had we used just one ``narrower'' command.\par} \endexample\enddesc \begindesc \cts leftskip {\param{glue}} \cts rightskip {\param{glue}} \explain These parameters tell \TeX\ how much glue to place at the left and at the right end of each line of the current paragraph. We'll just explain how |\leftskip| works since |\rightskip| is analogous. ^^{indentation} You can increase the left margin by setting |\leftskip| to a fixed nonzero \minref{dimension}. If you give |\leftskip| some stretch, you can produce ^{ragged left} text, i.e., text that has an uneven left margin. Ordinarily, you should enclose any \minref{assignment} to |\leftskip| in a \minref{group} together with the affected text in order to keep its effect from continuing to the end of your document. However, it's pointless to change |\leftskip|'s value inside a group that is in turn contained within a paragraph---the value of |\leftskip| at the \emph{end} of a paragraph is what determines how \TeX\ breaks the paragraph into lines. \minrefs{line break} \example {\leftskip = 1in The White Rabbit trotted slowly back again, looking anxiously about as it went, as if it had lost something. {\leftskip = 10in % has no effect It muttered to itself, ``The Duchess!! The Duchess!! She'll get me executed as sure as ferrets are ferrets!!''}\par}% | \produces {\leftskip = 1in The White Rabbit trotted slowly back again, looking anxiously about as it went, as if it had lost something. {\leftskip = 10in % has no effect It muttered to itself, ``The Duchess! The Duchess! She'll get me executed as sure as ferrets are ferrets!''}\par}% \nextexample \pretolerance = 10000 % Don't hyphenate. \rightskip = .5in plus 2em The White Rabbit trotted slowly back again, looking anxiously about as it went, as if it had lost something. It muttered to itself, ``The Duchess!! The Duchess!! She'll get me executed as sure as ferrets are ferrets!!'' | \produces \pretolerance = 10000 % Don't hyphenate. \rightskip = .5in plus 2em The White Rabbit trotted slowly back again, looking anxiously about as it went, as if it had lost something. It muttered to itself, ``The Duchess! The Duchess! She'll get me executed as sure as ferrets are ferrets!'' \endexample \enddesc \begindesc \easy\cts raggedright {} \cts ttraggedright {} \explain These commands cause \TeX\ to typeset your document ``^{ragged right}''. Interword spaces all have their natural size, i.e., they all have the same width and don't stretch or shrink. Consequently the right margin is generally not even. The alternative, which is \TeX's default, is to typeset your document justified, ^^{justification} i.e., with uniform left and right margins. In justified text, interword spaces are stretched in order to make the right margin even. Some typographers prefer ragged right because it avoids distracting ``rivers'' of white space on the printed page. \minrefs{justified text} You should use the |\ttraggedright| command when typesetting text in a monospaced font and the |\raggedright| command when typesetting text in any other font. Most of the time you'll want to apply these commands to an entire document, but you can limit their effects by enclosing them in a \minref{group}. \example \raggedright ``You couldn't have it if you {\it did\/} want it,'' the Queen said. ``The rule is, jam tomorrow and jam yesterday---but never jam {\it today\/}.'' ``It {\it must\/} come sometimes to `jam today,% thinspace'' Alice objected. ``No, it can't'', said the Queen. ``It's jam every {\it other\/} day: today isn't any {\it other\/} day.'' | \produces \raggedright ``You couldn't have it if you {\it did\/} want it,'' the Queen said. ``The rule is, jam tomorrow and jam yesterday---but never jam {\it today\/}.'' ``It {\it must\/} come sometimes to `jam today,% '\thinspace'' Alice objected. ``No, it can't'', said the Queen. ``It's jam every {\it other\/} day: today isn't any {\it other\/} day.'' \endexample \enddesc \begindesc \cts hang {} \explain This command indents the second and subsequent lines of a paragraph by |\parindent|, the paragraph ^{indentation} (\xref{\parindent}). Since the first line is already indented by |\parindent| (unless you've cancelled the indentation with |\noindent|), the entire paragraph appears to be indented by |\parindent|. \example \parindent=24pt \hang ``I said you {\it looked} like an egg, Sir,'' Alice gently explained to Humpty Dumpty. ``And some eggs are very pretty, you know,'' she added. | \produces \parindent=24pt \hang ``I said you {\it looked} like an egg, Sir,'' Alice gently explained to Humpty Dumpty. ``And some eggs are very pretty, you know,'' she added. \endexample \enddesc \begindesc \cts hangafter {\param{number}} \cts hangindent {\param{dimen}} \explain These two \minref{parameter}s jointly specify ``^{hanging indentation}'' for a paragraph. The hanging indentation indicates to \TeX\ that certain lines of the paragraph should be indented and the remaining lines should have their normal width. ^^{indentation} |\hangafter| determines which lines are indented, while |\hangindent| determines the amount of indentation and whether it occurs on the left or on the right: \ulist \li Let $n$ be the value of |\hangafter|. If $n < 0$, the first $-n$ lines of the paragraph will be indented. If $n\ge0$, all but the first $n$ lines of the paragraph will be indented. \li Let $x$ be the value of |\hangindent|. If $x\ge0$, the lines will be indented by $x$ on the left. If $x<0$ the lines will be indented by $-x$ on the right. \endulist When you specify hanging indentation, it applies only to the next paragraph (if you're in vertical mode) or to the current paragraph (if you're in horizontal mode). \TeX\ uses the values of |\hangafter| and |\hangindent| at the end of a paragraph, when it breaks that paragraph into lines.\minrefs{line break} Unlike most of the other paragraph-shaping parameters, |\hangafter| and |\hangindent| are reset to their default values at the start of each paragraph, namely, $1$ for |\hangafter| and $0$ for |\hangindent|. If you want to typeset a sequence of paragraphs with hanging indentation, use |\everypar| (\xref{\everypar}). ^^|\everypar//for hanging indentation| If you specify |\hangafter| and |\hangindent| as well as ^|\parshape|, \TeX\ ignores the |\hangafter| and |\hangindent|. \example \hangindent=6pc \hangafter=-2 This is an example of a paragraph with hanging indentation. In this case, the first two lines are indented on the left, but after that we return to unindented text. | \produces \hangindent=6pc \hangafter=-2 This is an example of a paragraph with hanging indentation. In this case, the first two lines are indented on the left, but after that we return to unindented text. \nextexample \hangindent=-6pc \hangafter=1 This is another example of a paragraph with hanging indentation. Here, all lines after the first have been indented on the right. The first line, on the other hand, has been left unindented. | \produces \hangindent=-6pc \hangafter=1 This is another example of a paragraph with hanging indentation. Here, all lines after the first have been indented on the right. The first line, on the other hand, has been left unindented. \endexample \enddesc \margin{{\tt\\textindent} has been moved to earlier in this section.} \begindesc \cts parshape {$n\; i_1 l_1\; i_2 l_2\; \ldots \;i_n l_n$} \explain This command specifies the shape of the first $n$ lines of a paragraph--- the next paragraph if you're in vertical mode and the current paragraph if you're in horizontal mode. The $i$'s and $l$'s are all dimensions. The first line is indented by $i_1$ and has length $l_1$, the second line is indented by $i_2$ and has length $l_2$, and so forth. If the paragraph has more than $n$ lines, the last indentation\slash length pair is used for the extra lines. To achieve special effects such as the one shown here, you usually have to experiment a lot, insert kerns here and there, and choose your words to fit the shape. |\parshape|, like ^|\hangafter| and ^|\hangindent|, is effective only for one paragraph. If you specify |\hangafter| and |\hangindent| as well as |\par!-shape|, \TeX\ ignores the ^|\hangafter| and ^|\hangindent|. By the way, the following example saves and restores |\fontdimen| values explicitly, using temporary registers, since |\fontdimen| changes are always global (see \xref\fontdimen). \ifodd\pageno\vfill\eject\fi % so the wineglass is on a single page. \example % A small font and close interline spacing make this work \smallskip\font\sixrm=cmr6 \sixrm \baselineskip=7pt \dimen0=\fontdimen3\font \dimen2=\fontdimen4\font \fontdimen3\font=1.8pt \fontdimen4\font=.9pt \noindent \hfuzz=.1pt \parshape 30 0pt 120pt 1pt 118pt 2pt 116pt 4pt 112pt 6pt 108pt 9pt 102pt 12pt 96pt 15pt 90pt 19pt 84pt 23pt 77pt 27pt 68pt 30.5pt 60pt 35pt 52pt 39pt 45pt 43pt 36pt 48pt 27pt 51.5pt 21pt 53pt 16.75pt 53pt 16.75pt 53pt 16.75pt 53pt 16.75pt 53pt 16.75pt 53pt 16.75pt 53pt 16.75pt 53pt 16.75pt 53pt 14.6pt 48pt 24pt 45pt 30.67pt 36.5pt 51pt 23pt 76.3pt The wines of France and California may be the best known, but they are not the only fine wines. Spanish wines are often underestimated, and quite old ones may be available at reasonable prices. For Spanish wines the vintage is not so critical, but the climate of the Bordeaux region varies greatly from year to year. Some vintages are not as good as others, so these years ought to be s\kern -.1pt p\kern -.1pt e\kern -.1pt c\hfil ially n\kern .1pt o\kern .1pt t\kern .1pt e\kern .1pt d\hfil: 1962, 1964, 1966. 1958, 1959, 1960, 1961, 1964, 1966 are also good California vintages. Good luck finding them!! \fontdimen3\font=\dimen0 \fontdimen4\font=\dimen2 | %\margin{Wineglass text replaced because of permissions problem.} \produces % A small font and close interline spacing make this work \smallskip\font\sixrm=cmr6 \sixrm \baselineskip=7pt \dimen0=\fontdimen3\font \dimen2=\fontdimen4\font \fontdimen3\font=1.8pt \fontdimen4\font=0.9pt \noindent \hfuzz=0.1pt % reordered to save a line \parshape 30 0pt 120pt 1pt 118pt 2pt 116pt 4pt 112pt 6pt 108pt 9pt 102pt 12pt 96pt 15pt 90pt 19pt 84pt 23pt 77pt 27pt 68pt 30.5pt 60pt 35pt 52pt 39pt 45pt 43pt 36pt 48pt 27pt 51.5pt 21pt 53pt 16.75pt 53pt 16.75pt 53pt 16.75pt 53pt 16.75pt 53pt 16.75pt 53pt 16.75pt 53pt 16.75pt 53pt 16.75pt 53pt 14.6pt 48pt 24pt 45pt 30.67pt 36.5pt 51pt 23pt 76.3pt The wines of France and California may be the best known, but they are not the only fine wines. Spanish wines are often underestimated, and quite old ones may be available at reasonable prices. For Spanish wines the vintage is not so critical, but the climate of the Bordeaux region varies greatly from year to year. Some vintages are not as good as others, so these years ought to be s\kern -.1pt p\kern -.1pt e\kern -.1pt c\hfil ially n\kern .1pt o\kern .1pt t\kern .1pt e\kern .1pt d\hfil: 1962, 1964, 1966. 1958, 1959, 1960, 1961, 1964, 1966 are also good California vintages. Good luck finding them! \fontdimen3\font=\dimen0 \fontdimen4\font=\dimen2 \endexample \eix^^{indentation} \enddesc \begindesc \cts prevgraf {\param{number}} \explain In horizontal mode, this parameter specifies the number of lines in the paragraph so far; in vertical mode, it specifies the number of lines in the previous paragraph. \TeX\ only sets |\prevgraf| after it has finished breaking some text into lines, i.e., at a math display or at the end of a paragraph. See \knuth{page~103} for more details about it. \enddesc \begindesc \cts vadjust {\rqbraces{\}} \explain This command inserts the specified \ just after the output line containing the position where the command occurs. ^^{vertical lists//inserting in paragraphs} You can use it, for instance, to cause a page eject or to insert extra space after a certain line. \example Some of these words are \vadjust{\kern8pt\hrule} to be found above the line and others are to be found below it. | \produces Some of these words are \vadjust{\kern8pt \hbox to \hsize{\hfil\vbox{\advance\hsize by -\parindent \hrule width \hsize}}} to be found above the line and others are to be found below it. \endexample \enddesc \see |\parindent| (\xref\parindent), |\parskip| (\xref\parskip), |\everypar| (\xref\everypar). \eix^^{line breaks//and paragraph shape} \eix^^{paragraphs//shaping} %========================================================================== \section {Line breaks} %========================================================================== \subsection {Encouraging or discouraging line breaks} \begindesc \bix^^{line breaks} \bix^^{line breaks//encouraging or discouraging} \ctspecial break {} \xrdef{hbreak} \explain This command forces a line break. Unless you do something to fill out the line, you're likely to get an ``underfull hbox'' complaint. |\break| can also be used in vertical mode. \example Fill out this line\hfil\break and start another one.\par % Use \hfil here to fill out the line. This line is underfull---we ended it\break prematurely. % This line causes an `underfull hbox' complaint. | \produces \hbadness = 10000 % avoid hbadness message Fill out this line\hfil\break and start another one.\par % Use \hfil here to fill out the line. This line is underfull---we ended it\break prematurely. % This line causes an `underfull hbox' complaint. \endexample\enddesc \begindesc \ctspecial nobreak {} \xrdef{hnobreak} \explain This command prevents a line break where it otherwise might occur. |\nobreak| can also be used in vertical mode. \example Sometimes you'll encounter a situation where a certain space\nobreak\qquad must not get lost. | \produces Sometimes you'll encounter a situation where a certain space\nobreak\qquad must not get lost. \endexample \enddesc \begindesc \ctspecial allowbreak {} \xrdef{hallowbreak} \explain This command tells \TeX\ to allow a line break where one could not ordinarily occur. It's most often useful within a math formula, since \TeX\ is reluctant to break lines there. ^^{line breaks//in math formulas} |\allowbreak| can also be used in vertical mode. \example Under most circumstances we can state with some confidence that $2+2\allowbreak=4$, but skeptics may disagree. \par For such moronic automata, it is not difficult to analyze the input/\allowbreak output behavior in the limit. | \produces Under most circumstances we can state with some confidence that $2+2\allowbreak=4$, but skeptics may disagree. \par For such moronic automata, it is not difficult to analyze the input/\allowbreak output behavior in the limit. \endexample\enddesc \begindesc \ctspecial penalty {\} \xrdef{hpenalty} \explain This command produces a \minref{penalty} item. The penalty item makes \TeX\ more or less willing to break a line at the point where that item occurs. A negative penalty, i.e., a bonus, encourages a line break; a positive penalty discourages a line break. A penalty of $10000$ or more prevents a break altogether, while a penalty of $-10000$ or less forces a break. |\penalty| can also be used in vertical mode. \secondprinting{\vfill\eject} \example \def\break{\penalty -10000 } % as in plain TeX \def\nobreak{\penalty 10000 } % as in plain TeX \def\allowbreak{\penalty 0 } % as in plain TeX | \endexample \enddesc \secondprinting{\vglue-\baselineskip\vskip0pt} \begindesc \cts obeylines {} \explain \TeX\ normally treats an end of line as a space. |\obeylines| instructs \TeX\ to treat each end of line as an end of paragraph, thus forcing a line break. |\obeylines| is often useful when you're typesetting verse or computer programs. ^^{verse, typesetting}^^{poetry, typesetting}^^{computer programs, typesetting} If any of your lines are longer than the effective line length (|\hsize|\tminus|\parindent|), however, you may get an extra line break within those lines. Because \TeX\ inserts the |\parskip| glue (\xref\parskip) between lines controlled by |\obeylines| (since it thinks each line is a paragraph), you should normally set |\parskip| to zero when you're using |\obeylines|. You can use the ^|\obeyspaces| command (\xref{\obeyspaces}) to get \TeX\ to take spaces within a line literally. |\obeylines| and |\obeyspaces| are often used together. \example \obeylines ``Beware the Jabberwock, my son!! \quad The jaws that bite, the claws that catch!! Beware the Jubjub bird, and shun \quad The frumious Bandersnatch!!'' | \produces \obeylines ``Beware the Jabberwock, my son! \quad The jaws that bite, the claws that catch! Beware the Jubjub bird, and shun \quad The frumious Bandersnatch!'' \endexample \enddesc \secondprinting{\vglue-\baselineskip\vskip0pt} \begindesc \easy\cts slash {} \explain This command produces a ^{solidus} (/) and also tells \TeX\ that it can break the line after the solidus, if necessary. \example Her oldest cat, while apparently friendly to most people, had a Jekyll\slash Hyde personality when it came to mice. | \produces Her oldest cat, while apparently friendly to most people, had a Jekyll\slash Hyde personality when it came to mice. \endexample \eix^^{line breaks//encouraging or discouraging} \enddesc \secondprinting{\vfill\eject} %========================================================================== \subsection {Line breaking parameters} \begindesc \bix^^{line breaks//parameters affecting} % \cts pretolerance {\param{number}} \cts tolerance {\param{number}} \explain These parameters determine the \minref{badness} that \TeX\ will tolerate on each line when it is choosing line breaks for a paragraph. The badness is a measure of how far the interword spacing deviates from the ideal. |\pretolerance| specifies the tolerable badness for line breaks without hyphenation; |\tolerance| specifies the tolerable badness for line breaks with hyphenation. The tolerable badness can be exceeded in either of two ways: a line is too tight (the interword spaces are too small) or it is too loose (the interword spaces are too big). \ulist \li If \TeX\ must set a line too loosely, it complains about an ``underfull hbox''. \li If \TeX\ must set a line too rightly, it lets the line run past the right margin and complains about an ``overfull \minref{hbox}''. \endulist \noindent \TeX\ chooses line breaks in the following steps: \olist \li It attempts to choose line breaks without hyphenating. If none of the resulting lines have a badness exceeding |\pretolerance|, the line breaks are acceptable and the paragraph can now be set. \li Otherwise, it tries another set of line breaks, this time allowing hyphenation. If none of the resulting lines have a badness exceeding |\tolerance|, the new set of line breaks is acceptable and the paragraph can now be set. \li Otherwise, it adds ^|\emergencystretch| (see below) to the stretch of each line and tries again. \li If none of these attempts have produced an acceptable set of line breaks, it sets the paragraph with one or more overfull hboxes and complains about them. \endolist \PlainTeX\ sets |\tolerance| to $200$ and |\pretolerance| to $100$. If you set |\tolerance| to $10000$, \TeX\ becomes infinitely tolerant and accepts any spacing, no matter how bad (unless it encounters a word that won't fit on a line, even with hyphenation). Thus by changing |\tolerance| you can avoid overfull and underfull hboxes, but at the cost of making the spacing worse. By making |\pretolerance| larger you can get \TeX\ to avoid hyphenation (and also run faster), again at the cost of possibly worse spacing. If you set |\pretolerance| to $-1$, \TeX\ will not even try to set the paragraph without hyphenation. The ^|\hbadness| parameter (\xref \hbadness) determines the level of badness that \TeX\ will tolerate before it complains, but |\hbadness| does not affect the way that \TeX\ typesets your document. The ^|\hfuzz| parameter (\xref \hfuzz) determines the amount that an hbox can exceed its specified width before \TeX\ considers it to be erroneous. \enddesc \begindesc \cts emergencystretch {\param{dimen}} \explain By setting this parameter to be greater than zero, you can make it easier for \TeX\ to typeset your document without generating overfull hboxes. ^^{overfull boxes} This is a better alternative than setting |\tolerance=10000|, since that tends to produce really ugly lines. If \TeX\ can't typeset a paragraph without exceeding ^|\tolerance|, it will try again, adding |\emergencystretch| to the stretch of each line. The effect of the change is to scale down the badness of each line, enabling \TeX\ to make spaces wider than they would otherwise be and thus choose line breaks that are as good as possible under the circumstances. \enddesc \begindesc \cts looseness {\param{number}} \explain \minrefs{line break} This parameter gives you a way to change the total number of lines in a paragraph from what they optimally would be. |\looseness| is so named because it's a measure of how loose the paragraph is, i.e., how much extra space there is in it. Normally, |\looseness| is $0$ and \TeX\ chooses line breaks in its usual way. But if |\looseness| is, say, $3$, \TeX\ does the following: \olist \li It chooses line breaks normally, resulting in a paragraph of $n$ lines. \li It discards these line breaks and tries to find a new set of line breaks that gives the paragraph $n+3$ lines. (Without the previous step, \TeX\ wouldn't know the value of $n$.) \li If the previous attempt results in lines whose badness exceeds |\tol!-er!-ance|, ^^|\tolerance| it tries to get $n+2$ lines---and if that also fails, $n+1$ lines, and finally $n$ lines again. \endolist \noindent Similarly, if looseness is $-n$, \TeX\ attempts to set the paragraph with $n$ fewer lines than normal. The easiest way for \TeX\ to make a paragraph one line longer is to put a single word on the excess line. You can prevent this by putting a tie (\xref{@not}) between the last two words of the paragraph. Setting |\looseness| is the best way to force a paragraph to occupy a given number of lines. Setting it to a negative value is useful when you're trying to increase the amount of text you can fit on a page. Similarly, setting it to a positive value is useful when you're trying to decrease the amount of text on a page. \TeX\ sets |\looseness| to $0$ when it ends a paragraph, after breaking the paragraph into lines. If you want to change the looseness of several paragraphs, you must do it individually for each one or put the change into |\everypar| \ctsref\everypar. ^^|\everypar//for setting \b\tt\\looseness\e| \enddesc \begindesc \cts linepenalty {\param{number}} \explain \minrefs{line break} This parameter specifies the \minref{penalty} that \TeX\ assesses for each line break when it is breaking a paragraph into lines. The penalty is independent of where the line break occurs. Increasing the value of this parameter causes \TeX\ to try harder to set a paragraph with a minimum number of lines, even at the cost of other aesthetic considerations such as avoiding overly tight interword spacing. Demerits are in units of \minref{badness} squared, so you need to assign a rather large value to this parameter (in the thousands) for it to have any effect. \PlainTeX\ sets |\linepenalty| to $10$. \enddesc \begindesc \cts adjdemerits {\param{number}} \explain \minrefs{line break} ^^{hyphenation//penalties for} {\tighten This parameter specifies additional \minref{demerits} that \TeX\ attaches to a breakpoint between two adjacent lines that are ``visually incompatible''. Such a pair of lines makes a paragraph appear uneven. Incompatibility is evaluated in terms of the tightness or looseness of lines: } \olist\compact \li A line is tight if its \minref{glue} needs to shrink by at least $50\%$. \li A line is decent if its badness is $12$ or less. \li A line is loose if its glue needs to stretch by more than $50\%$. \li A line is very loose if its glue needs to stretch so much that its badness exceeds $100$. \endolist Two adjacent lines are visually incompatible if their categories are not adjacent, e.g., a tight line is next to a loose one or a decent line is next to a very loose one. Demerits are in units of \minref{badness} squared, so you need to assign a rather large value to this parameter (in the thousands) for it to have any effect. \PlainTeX\ sets |\adjdemerits| to~$10000$. \enddesc \begindesc \bix^^{hyphenation//penalties for} \cts exhyphenpenalty {\param{number}} \explain \minrefs{line break} This parameter specifies the \minref{penalty} that \TeX\ attaches to a breakpoint at an explicit hyphen such as the one in ``helter-skelter''. Increasing this parameter has the effect of discouraging \TeX\ from ending a line at an explicit hyphen. \PlainTeX\ sets |\exhyphenpenalty| to $50$. \enddesc \begindesc \cts hyphenpenalty {\param{number}} \explain \minrefs{line break} This parameter specifies the \minref{penalty} that \TeX\ attaches to a breakpoint at an implicit hyphen. Implicit hyphens can come from \TeX's hyphenation dictionary or from ^{discretionary hyphens} that you've inserted with |\-|~(\xref{\@minus}). ^^|-//leads to {\tt\\hyphenpenalty}| Increasing this parameter has the effect of discouraging \TeX\ from hyphenating words. \PlainTeX\ sets |\hyphenpenalty| to $50$. \enddesc \begindesc \cts doublehyphendemerits {\param{number}} \explain \minrefs{line break} {\tighten This parameter specifies additional \minref{demerits} that \TeX\ attaches to a breakpoint when that breakpoint leads to two consecutive lines that end in a hyphen. Increasing the value of this parameter has the effect of discouraging \TeX\ from hyphenating two lines in a row. Demerits are in units of \minref{badness} squared, so you need to assign a rather large value to this parameter (in the thousands) for it to have any effect. \PlainTeX\ sets |\doublehyphendemerits| to $10000$. } \enddesc \begindesc \cts finalhyphendemerits {\param{number}} \explain \minrefs{line break} {\tighten This parameter specifies additional \minref{demerits} that \TeX\ attaches to a breakpoint that causes the next to last line of a paragraph to end with a hyphen. Such a hyphen is generally considered to be unaesthetic because of the possible blank space from a short last line beneath it. Increasing the value of this parameter has the effect of discouraging \TeX\ from ending the next to the last line with a hyphen. Demerits are in units of \minref{badness} squared, so you need to assign a rather large value to this parameter (in the thousands) for it to have any effect. \PlainTeX\ sets |\finalhyphendemerits| to $5000$. } \eix^^{hyphenation//penalties for} \enddesc \begindesc \cts binoppenalty {\param{number}} \explain ^^{operators} This parameter specifies the penalty for breaking a math formula after a binary operator when the formula appears in a paragraph. \PlainTeX\ sets |\binoppenalty| to $700$. \enddesc \begindesc \cts relpenalty {\param{number}} \explain ^^{relations} This parameter specifies the penalty for breaking a math formula after a relation when the formula appears in a paragraph. \PlainTeX\ sets |\rel!-penal!-ty| to~$500$. \eix^^{line breaks//parameters affecting} \enddesc %========================================================================== \subsection {Hyphenation} \begindesc \bix^^{hyphenation} % \easy\ctspecial - \ctsxrdef{@minus} \explain The |\-| command inserts a ``discretionary hyphen'' ^^{discretionary hyphens} into a word. The discretionary hyphen allows \TeX\ to hyphenate the word at that place. \TeX\ isn't obliged to hyphenate there---it does so only if it needs to. This command is useful when a word that occurs in one or two places in your document needs to be hyphenated, but \TeX\ can't find an appropriate hyphenation point on its own. \example Alice was exceedingly reluctant to shake hands first with either Twee\-dle\-dum or Twee\-dle\-dee, for fear of hurting the other one's feelings. | \produces Alice was exceedingly reluctant to shake hands first with either Twee\-dle\-dum or Twee\-dle\-dee, for fear of hurting the other one's feelings. \endexample \enddesc \begindesc \cts discretionary {\rqbraces{\} \rqbraces{\} \rqbraces{\}} \explain \minrefs{line break} ^^{hyphenation} This command specifies a ``discretionary break'', namely, a place where \TeX\ can break a line. It also tells \TeX\ what text to put on either side of the break. \ulist \li If \TeX\ does not break there, it uses the \. \li If \TeX\ does break there, it puts the \ just before the break and the \ just after the break. \endulist \noindent Just as with |\-|, \TeX\ isn't obligated to break a line at a discretionary break. In fact, |\-| is ordinarily equivalent to |\discretionary!allowbreak{-}{}{}|. \TeX\ sometimes inserts discretionary breaks on its own. For example, it inserts |\discretionary!allowbreak{}{}{}| after an explicit hyphen or dash. {\hyphenchar\tentt=-1 % needed to avoid weirdnesses \example % An ordinary discretionary hyphen (equivalent to \-): \discretionary{-}{}{} % A place where TeX can break a line, but should not % insert a space if the line isn't broken there, e.g., % after a dash: \discretionary{}{}{} % Accounts for German usage: `flicken', but `flik- % ken': German ``fli\discretionary{k-}{k}{ck}en'' | ^^{hyphenation//German} \endexample} \enddesc \begindesc \cts hyphenation {\rqbraces{\\thinspace\vs\ $\ldots$\ \vs \thinspace\}} \explain \TeX\ keeps a dictionary of exceptions to its ^{hyphenation} rules. Each dictionary entry indicates how a particular word should be hyphenated. The |\hyphenation| command adds words to the dictionary. Its argument is a sequence of words separated by blanks. Uppercase and lowercase letters are equivalent. The hyphens in each word indicate the places where \TeX\ can hyphenate that word. A word with no hyphens in it will never be hyphenated. However, you can still override the hyphenation dictionary by using |\-| in a particular occurrence of a word. You need to provide all the grammatical forms of a word that you want \TeX\ to handle, e.g., both the singular and the plural. \example \hyphenation{Gry-phon my-co-phagy} \hyphenation{man-u-script man-u-scripts piz-za} | \endexample \enddesc \begindesc \cts uchyph {\param{number}} \explain A positive value of |\uchyph| (uppercase hyphenation) permits hyphenation of words, such as proper names, that start with a capital letter. A zero or negative value inhibits such hyphenation. \PlainTeX\ sets |\uchyph| to $1$, so \TeX\ normally tries to hyphenate words that start with a capital letter. \enddesc \begindesc \cts showhyphens {\rqbraces{\\thinspace\vs\ $\ldots$\ \vs \thinspace\}} \explain This command isn't normally used in documents, but you can use it at your terminal to see how \TeX\ would hyphenate some random set of words. The words, with hyphenations indicated, appear both in the log and at your terminal. You'll get a complaint about an underfull hbox---just ignore it. \example \showhyphens{threshold quizzical draughts argumentative} | \logproduces Underfull \hbox (badness 10000) detected at line 0 [] \tenrm thresh-old quizzi-cal draughts ar-gu-men-ta-tive | \endexample \enddesc \begindesc \cts language {\param{number}} \explain Different languages have different sets of hyphenation rules. This parameter determines the set of ^{hyphenation rules} that \TeX\ uses. By changing |\language| you can get \TeX\ to hyphenate portions of text or entire documents according to the hyphenation rules appropriate to a particular language. ^^{European languages} Your ^{local information} about \TeX\ will tell you if any additional sets of hyphenation rules are available (besides the ones for English) and what the appropriate values of |\language| are. The default value of |\language| is $0$. \TeX\ sets the current language to $0$ at the start of every paragraph, and compares |\language| to the current language whenever it adds a character to the current paragraph. If they are not the same, \TeX\ adds a ^{whatsit} indicating the language change. This whatsit is the clue in later processing that the language rules should change. \enddesc \begindesc \cts setlanguage {\} \explain This command sets the current language to \ by inserting the same whatsit that you'd get by changing ^|\language|. However, it does not change the value of |\language|. \enddesc \begindesc \cts lefthyphenmin {\param{number}} \cts righthyphenmin {\param{number}} \explain These parameters specify the smallest word fragments that \TeX\ allows at the left and at the right end of a hyphenated word. \PlainTeX\ defaults them to $2$ and $3$ respectively; these are the recommended values for English. \enddesc \begindesc \bix^^{fonts//hyphenation characters for} \cts hyphenchar {\\param{number}} \explain \TeX\ doesn't necessarily use the `-' character at hyphenation points. Instead, it uses the |\hyphenchar| of the current font, which is usually `-' but need not be. If a font has a negative |\hyphenchar| value, \TeX\ won't hyphenate words in that font. Note that \ is a control sequence that names a font, not a \ that names font files. Beware: an assignment to |\hyphenchar| is \emph{not} undone at the end of a group. If you want to change |\hyphenchar| locally, you'll need to save and restore its original value explicitly. \example \hyphenchar\tenrm = `- % Set hyphenation for tenrm font to `-'. \hyphenchar\tentt = -1 % Don't hyphenate words in font tentt. | \endexample \enddesc \begindesc \cts defaulthyphenchar {\param{number}} \explain When \TeX\ reads the metrics file ^^{metrics file//default hyphen in} for a font in response to a ^|\font| command, it sets the font's ^|\hyphenchar| to |\default!-hyphen!-char|. If the value of |\default!-hyphen!-char| is not in the range $0$--$255$ when you load a font, \TeX\ won't hyphenate any words in that font unless you override the decision by setting the font's |\hyphenchar| later on. \PlainTeX\ sets |\default!-hyphen!-char| to $45$, the \ascii\ code for `|-|'. \example \defaulthyphenchar = `- % Assume `-' is the hyphen, unless overridden. \defaulthyphenchar = -1 % Don't hyphenate, unless overridden. | \endexample \eix^^{fonts//hyphenation characters for} \enddesc \see |\pretolerance| (\xref \pretolerance). \eix^^{hyphenation} \eix^^{line breaks} %========================================================================== \section {Section headings, lists, and theorems} \begindesc ^^{section headings} \easy\ctspecial beginsection {\\thinspace{\bt\\par}} \ctsxrdef{@beginsection} \explain You can use this command to begin a major subdivision of your document. \ is intended to serve as a section title. |\beginsection| surrounds \ by extra vertical space and sets it in boldface, left-justified. You can produce the |\par| that ends \ with a blank line. \let\message = \gobble % Don't bother to tell us about Pig and Pepper. \example $\ldots$ till she had brought herself down to nine inches high. \beginsection Section 6. Pig and Pepper For a minute or two she stood looking at the house $\ldots$ | \produces $\ldots$ till she had brought herself down to nine inches high. \beginsection Section 6. Pig and Pepper For a minute or two she stood looking at the house $\ldots$ \endexample \enddesc \begindesc \cts item {\} \cts itemitem {\} \explain ^^{itemized lists} These commands are useful for creating ^{itemized lists}. The entire paragraph following \ is indented by |\parindent| ^^|\parindent//indentation for itemized lists| (for |\item|) or by |2\parindent| (for |\itemitem|). (See \xrefpg{\parindent} for an explanation of |\parindent|.) Then \, followed by an en space, is placed just to the left of the text of the first line of the paragraph so that it falls within the paragraph indentation as specified by |\parindent|. If you want to include more than one paragraph in an item, put |\item{}| in front of the additional paragraphs. \example {\parindent = 18pt \noindent Here is what we require: \item{1.}Three eggs in their shells, but with the yolks removed. \item{2.}Two separate glass cups containing: \itemitem{(a)}One-half cup {\it used} motor oil. \itemitem{(b)}One cup port wine, preferably French. \item{3.}Juice and skin of one turnip.} | \produces {\parindent = 18pt \noindent Here is what we require: \item{1.}Three eggs in their shells, but with the yolks removed. \item{2.}Two separate glass cups containing: \itemitem{(a)}One-half cup {\it used} motor oil. \itemitem{(b)}One cup port wine, preferably French. \item{3.}Juice and skin of one turnip.} \endexample \enddesc \begindesc \easy\ctspecial proclaim {\{\tt.}\vs\thinspace \\thinspace{\bt\\par}} \ctsxrdef{@proclaim} \explain ^^{theorems} ^^{lemmas} ^^{hypotheses} This command ``proclaims'' a theorem, lemma, hypothesis, etc. It sets \ in boldface type and the following paragraph in italics. \ must be followed by a period and a space token, which serve to set off \ from \. \ consists of the text up to the next paragraph boundary. You can include multiple paragraphs by using |\endgraf| instead of a blank line or |\par|. \example \proclaim Theorem 1. What I say is not to be believed. \proclaim Corollary 1. Theorem 1 is false.\par | \produces \proclaim Theorem 1. What I say is not to be believed. \proclaim Corollary 1. Theorem 1 is false.\par \endexample \enddesc \enddescriptions \endchapter \byebye