top a b c d e f g h i j k l m n o p r s t u v w x y (family order)

top a b c d e f g h i j k l m n o p r s t u v w x y (family order)

top a b c d e f g h i j k l m n o p r s t u v w x y (family order)

top a b c d e f g h i j k l m n o p r s t u v w x y (family order)

top a b c d e f g h i j k l m n o p r s t u v w x y (family order)

top a b c d e f g h i j k l m n o p r s t u v w x y (family order)

top a b c d e f g h i j k l m n o p r s t u v w x y (family order)

top a b c d e f g h i j k l m n o p r s t u v w x y (family order)

top a b c d e f g h i j k l m n o p r s t u v w x y (family order)

top a b c d e f g h i j k l m n o p r s t u v w x y (family order)

top a b c d e f g h i j k l m n o p r s t u v w x y (family order)

top a b c d e f g h i j k l m n o p r s t u v w x y (family order)

top a b c d e f g h i j k l m n o p r s t u v w x y (family order)

top a b c d e f g h i j k l m n o p r s t u v w x y (family order)

top a b c d e f g h i j k l m n o p r s t u v w x y (family order)

top a b c d e f g h i j k l m n o p r s t u v w x y (family order)

top a b c d e f g h i j k l m n o p r s t u v w x y (family order)

top a b c d e f g h i j k l m n o p r s t u v w x y (family order)

top a b c d e f g h i j k l m n o p r s t u v w x y (family order)

top a b c d e f g h i j k l m n o p r s t u v w x y (family order)

top a b c d e f g h i j k l m n o p r s t u v w x y (family order)

top a b c d e f g h i j k l m n o p r s t u v w x y (family order)

top a b c d e f g h i j k l m n o p r s t u v w x y (family order)

top a b c d e f g h i j k l m n o p r s t u v w x y (family order)

top a b c d e f g h i j k l m n o p r s t u v w x y (family order)

top a b c d e f g h i j k l m n o p r s t u v w x y (family order)

top Box Character Debugging File I/O Fonts Glue Hyphenation Inserts Job Kern Logic Macro Marks Math Page Paragraph Penalties Registers Tables (alpha order)

top Box Character Debugging File I/O Fonts Glue Hyphenation Inserts Job Kern Logic Macro Marks Math Page Paragraph Penalties Registers Tables (alpha order)

top Box Character Debugging File I/O Fonts Glue Hyphenation Inserts Job Kern Logic Macro Marks Math Page Paragraph Penalties Registers Tables (alpha order)

top Box Character Debugging File I/O Fonts Glue Hyphenation Inserts Job Kern Logic Macro Marks Math Page Paragraph Penalties Registers Tables (alpha order)

top Box Character Debugging File I/O Fonts Glue Hyphenation Inserts Job Kern Logic Macro Marks Math Page Paragraph Penalties Registers Tables (alpha order)

top Box Character Debugging File I/O Fonts Glue Hyphenation Inserts Job Kern Logic Macro Marks Math Page Paragraph Penalties Registers Tables (alpha order)

top Box Character Debugging File I/O Fonts Glue Hyphenation Inserts Job Kern Logic Macro Marks Math Page Paragraph Penalties Registers Tables (alpha order)

top Box Character Debugging File I/O Fonts Glue Hyphenation Inserts Job Kern Logic Macro Marks Math Page Paragraph Penalties Registers Tables (alpha order)

top Box Character Debugging File I/O Fonts Glue Hyphenation Inserts Job Kern Logic Macro Marks Math Page Paragraph Penalties Registers Tables (alpha order)

top Box Character Debugging File I/O Fonts Glue Hyphenation Inserts Job Kern Logic Macro Marks Math Page Paragraph Penalties Registers Tables (alpha order)

top Box Character Debugging File I/O Fonts Glue Hyphenation Inserts Job Kern Logic Macro Marks Math Page Paragraph Penalties Registers Tables (alpha order)

top Box Character Debugging File I/O Fonts Glue Hyphenation Inserts Job Kern Logic Macro Marks Math Page Paragraph Penalties Registers Tables (alpha order)

top Box Character Debugging File I/O Fonts Glue Hyphenation Inserts Job Kern Logic Macro Marks Math Page Paragraph Penalties Registers Tables (alpha order)

top Box Character Debugging File I/O Fonts Glue Hyphenation Inserts Job Kern Logic Macro Marks Math Page Paragraph Penalties Registers Tables (alpha order)

top Box Character Debugging File I/O Fonts Glue Hyphenation Inserts Job Kern Logic Macro Marks Math Page Paragraph Penalties Registers Tables (alpha order)

top Box Character Debugging File I/O Fonts Glue Hyphenation Inserts Job Kern Logic Macro Marks Math Page Paragraph Penalties Registers Tables (alpha order)

top Box Character Debugging File I/O Fonts Glue Hyphenation Inserts Job Kern Logic Macro Marks Math Page Paragraph Penalties Registers Tables (alpha order)

top Box Character Debugging File I/O Fonts Glue Hyphenation Inserts Job Kern Logic Macro Marks Math Page Paragraph Penalties Registers Tables (alpha order)

top Box Character Debugging File I/O Fonts Glue Hyphenation Inserts Job Kern Logic Macro Marks Math Page Paragraph Penalties Registers Tables (alpha order)

top Box Character Debugging File I/O Fonts Glue Hyphenation Inserts Job Kern Logic Macro Marks Math Page Paragraph Penalties Registers Tables (alpha order)

Description:

A discretionary hyphen inserts `\discretionary{\charh}{}{}' into a horizontal or math list, provided that h, the \hyphenchar of the current font, is between 0 and 255. Otherwise, it inserts `\discretionary{}{}{}' into the list [455]. If the command occurs in vertical mode, it starts a new paragraph [283]. TeX can break a compound word at an existing hyphen, but it does not apply its hyphenation routine to such words. So, a discretionary hyphen may be required to prevent a bad line break. For example `mass-produced' might need to be changed to `mass-pro\-duced' [96].

  1. \def\tstoryA{There are cries, sobs, confusion among the people, and at
  2. that moment the cardinal himself, the Grand Inquisitor, passes by the
  3. cathedral. He is an old man . . .}
  4. \def\tstoryB{There are cries, sobs, confusion a\-mong the people, and
  5. at that moment the cardinal himself, the Grand Inquisitor, passes by
  6. the cathedral. He is an old man . . .}
  7. \setbox0=\vbox{\hsize=2.25in \tstoryA}
  8. \setbox1=\vbox{\hsize=2.25in \tstoryB}
  9. \hbox to \hsize{\box0\hfill\box1}

Comments:

• Line 4 adds a discretionary hyphen to `among'.

See Also: discretionary, hyphenchar.

Description:

EVERY character in every font has a dimension called the italic correction. For most characters and most fonts that dimension is zero. But fonts that slant (e.g., an italics font) may have many characters with a nonzero dimension. In horizontal mode `\/' means ``add an explicit kern using the italic correction for the preceeding character or ligature to the current list'' [14 and 287]. But in math mode `\/' adds a kern of width zero to the current list [292].

  1. \selffamily AA
  2.        [{\it f}]   [{\it f\/}]   [\/{\it f\/}]   {\it [}f]   {\it [\/}f\/]\par
  3. \bgroup
  4.     \font\rm=cmr12 scaled 2000
  5.     \font\it=cmti12 scaled 2000
  6.     \rm[{\it f}]   [{\it f\/}]   [\/{\it f\/}]   {\it [}f]   {\it [\/}f\/]\par
  7. \egroup\par
  8. [\R0{\it f\/}] [\R1{\it f\/}] [\R2{\it f\/}] [\R3{\it f\/}] [\R4{\it f\/}]
  9. [\R5{\it f\/}] [\R6{\it f\/}] [\R7{\it f\/}] [\R8{\it f\/}] [\R9{\it f\/}]
 10. %\def\R#1{\dimen0=0.05em\multiply\dimen0 by #1\kern\dimen0\relax}

Comments:

• This example uses roman and italics fonts for Caslon 224 and Computer Modern at 24 points.
• The first [f] (made by lines 2 and 6) shows an instance where the f\/ requires an italic correction. It is added in the second [f].
• TeX does not provide a pre-character italic correction, and the third [f] shows [ does not have an italic correction. This is not a problem for CM, but it is for Caslon. It represents something that will need to be lived with or fixed by hand [\R4f].
• The last two examples show the Caslon [ does not have an italic correction, but the CM one does.
• Lines 8-9 shows several examples of hand kerning. The author's preference uses \R4. The definition of \R is shown on line 10.
• One of the lessons here is that changing fonts can introduce subtle problems.

See Also: font.

Description:

IN certain situations TeX ignores spaces (e.g., after a control word). In other situations it converts multiple spaces into a single space. A control space is used to force TeX to typeset a space [8]. A control space is required when using \nonfrenchspacing to fix the spacing after periods which occur in the middle of a sentence [74]. Regular spaces and blank lines are ignored in vertical mode. However, a control space in vertical mode begins a new paragraph [87]. In math mode regular spaces are ignored, and a control space is required to separate words [163]. A control space adds the same glue to a horizontal or math list that a <space token> with space factor 1000 adds [285 and 290]. Plain TeX defines \<tab> and \<return> so they return a control space [351]. Verbatim listings often make the space character active (category 13) and set it equal to a control space [381].

     If there were no \]'s in \tex, using \TeX{} would be more difficult
     but not impossible.

Comments:

• TeX does not remove spaces following a `}'.

Description:

THIS command places the subformula preceeding \above over the subformula following \above. Then it places a rule whose thickness is <dimen> between the two subformulas. The command is a generalization of \over and \atop [143] and a special case of \abovewithdelims [152].

  1. \def\tabove#1%
  2. {%
  3.      {{2 \over 3}\above#1 {1 \over 6}} = 
  4.      {{2 \over 3}\cdot{6 \over 1}} = {12 \over 3} = 4
  5. }
  6. $$\hbox{$\tabove{1pt}$,}\quad
  7.   \hbox{$\tabove{2pt}$,}\quad
  8.   \tabove{1pt},\quad
  9.   \tabove{2pt}$$

Comments:

• Lines 1-5 define a macro which uses \above to typeset an expression that might appear in a fourth or fifth grade mathematics text. The macro has a single parameter which determines the thickness of the main fraction bar used in the first fraction.
• Lines 6-9 create a display and typeset the expression four times: twice in an \hbox in regular math mode and twice in display math mode.

See Also: over, atop, abovewithdelims.

Description:

AFTER TeX processes a math display, it adds the following items to the current vertical list: a penalty, glue, the actual display, glue, and a penalty. The two glue items are either the above and below displayskips or the above and below displayshortskips. The latter are used if: a) the display occurs at the start of a paragraph; b) the display follows another display; or c) the display occurs in the middle of a paragraph but does not have a left equation number and does not overlap the preceeding line of text [189].

     \def\tfn#1#2#3#4#5% above & below regular & short skips + post-example skip
     {%
        \abovedisplayskip=#1
        \belowdisplayskip=#2
        \abovedisplayshortskip=#3
        \belowdisplayshortskip=#4
        \hfill The Fibonacci numbers are defined\break 
        by: $F_0=1$, $F_1=1$, and
        $$F_n = F_{n-1} + F_{n-2} \hbox{, for } n \ge 2.$$ 
        A related sequence $\{v_n\}$ is: 2, 1, 3, 4, 7, 11, 18, \char144\ It
        satisfies $v_n$ = $x^n + y^n$, where $x = (1 + \sqrt 5)/ 2$ 
        and $y = (1 - \sqrt 5)/ 2$.
        \vskip#5
     }
     \tfn{12pt}{12pt}{0pt}{0pt}{1\baselineskip}
     \tfn{12pt}{12pt}{3pt}{3pt}{0pt}

Comments:

• This example is very similar to the one for \abovedisplayskip only this time each example uses short display skips.
• The examples illustrate the two extremes of the short skip Plain TeX provides.

See Also: belowdisplayshortskip, abovedisplayskip.

Description:

AFTER TeX processes a math display, it adds the following items to the current vertical list: a penalty, glue, the actual display, glue, and a penalty. The two glue items are either the above and below displayskips or the above and below displayshortskips. The former are used if: a) the display occurs in the middle of a paragraph and the display overlaps the preceeding line of text; or b) the display has a left equation number [189]. Also, the former are used before and after an alignment display [190].

     \def\tfn#1#2#3#4#5% above & below regular & short skips + post-example skip
     {%
        \abovedisplayskip=#1
        \belowdisplayskip=#2
        \abovedisplayshortskip=#3
        \belowdisplayshortskip=#4
        The Fibonacci numbers are defined by: $F_0=1$, $F_1=1$, and
        $$F_n = F_{n-1} + F_{n-2} \hbox{, for } n \ge 2.$$
        A related sequence $\{v_n\}$ is: 2, 1, 3, 4, 7, 11, 18, \char144\ It
        satisfies $v_n$ = $x^n + y^n$, where $x = (1 + \sqrt 5)/ 2$ 
        and $y = (1 - \sqrt 5)/ 2$.
        \vskip#5
     }
     \tfn{3pt}{3pt}{6pt}{6pt}{1\baselineskip}
     \tfn{12pt}{12pt}{6pt}{6pt}{0pt}

Comments:

• This example typesets the same material twice using regular display skips.
• The first example uses a small value for the glue placed before and after the display. It is what Plain TeX will make if the glue shrinks its maximum amount.
• The second example shows what Plain TeX will make if the glue neither stretches nor shrinks.

See Also: belowdisplayskip, abovedisplayshortskip.

Description:

THIS command is a generalized form of \above. It builds the same fraction \above does (with a line whose thickness is <dimen>). Then it puts <delim1> before the fraction and <delim2> after it. Both delimiters are scaled appropriately. If a `.' is used for either <delim>, TeX ignores the corresponding delimiter. In fact, \above<dimen> is an abbreviation for `\abovewithdelims..<dimen>'. All the other fraction commands are derived from this command [152].

  1. \def\legendre{\overwithdelims()}
  2. \def\tlA{\abovewithdelims()0.4pt}
  3. \def\tlB{\abovewithdelims()1.0pt}
  4. \def\tlC{\abovewithdelims()2.0pt}
  5. $a\legendre p$, ${a\tlA p}$, ${a\tlB p}$, and ${a\tlC p}$.
  6. $$\displaylines{{a\legendre p} = +1\hbox{,\quad if\quad}a\hbox{\ R }p,\qquad
  7.            {a\tlA p} = -1\hbox{,\quad if\quad}a\hbox{\ N }p,\qquad
  8.            {a\tlC p} = -1\hbox{,\quad if\quad}a\hbox{\ N }p.\cr}$$

Comments:

• Lines 1-4 define four macros which make a Legendre symbol. The first macro uses \overwithdelims, and each of the remaining three use \abovewithdelims.
• Line 5 uses each macro to typeset an example in math mode.
• Lines 6-8 typeset three of the macros in display mode.

See Also: above, over, atop, overwithdelims, atopwithdelims.

Description:

THIS command places an accent on a character. The <8-bit number> gives the position of the accent in the current font. The accent should be at the correct position for characters whose height is the x-height of the current font. The accent is raised or lowered to fit taller or shorter characters. The width of the result is the width of the character being accented [54]. The <optional assignments> stands for 0 or more assignments [275-278] other than \setbox. The character receiving the accent should be: a character with category code 11 or 12 (letter or other), a \char or \chardef token, or \noboundary. If a character does not follow the assignment, the result is the same as \char<8-bit number> [286]. Otherwise, the character is accented by the accent. After an accent TeX sets \spacefactor=1000 [287]. If \accent appears in vertical mode, a new paragraph is started [283]. Plain TeX sets up the basic accents for Computer Modern fonts [356]. If another font is used which has the accents in different positions, new definitions are required [nr].

  1. % Accents for 9y encoding of Caslon 224:
  2. \def\'#1{{\accent00 #1}}% acute
  3. \def\u#1{{\accent01 #1}}% breve
  4. \def\v#1{{\accent02 #1}}% hacek or check
  5. \def\^#1{{\accent04 #1}}% circumflex or hat
  6. \def\"#1{{\accent05 #1}}% umlaut
  7. \def\.#1{{\accent06 #1}}% dot (over)
  8. \def\`#1{{\accent07 #1}}% grave
  9. \def\H#1{{\accent08 #1}}% long Humgarian umlaut
 10. \def\=#1{{\accent09 #1}}% macron or bar
 11. \def\r#1{{\accent11 #1}}% ring (open dot over)
 12. \def\~#1{{\accent12 #1}}% tilde
 13. \def\d#1{{\o@lign{\relax#1\crcr\hidewidth\sh@ft{10}.\hidewidth}}}% dot (under)
 14. \def\b#1{{\o@lign{\relax#1\crcr\hidewidth\sh@ft{29}%               bar (under)
 15.     \vbox to.2ex{\hbox{\char09}\vss}\hidewidth}}}
 16. \def\c#1{\setbox\z@\hbox{#1}\dimen\z@=\dp\z@\ifdim\ht\z@=1ex\accent03 #1%cedilla
 17.   \else{\ooalign{\unhbox\z@\crcr\hidewidth\lower0.9\dimen\z@ %
 18.         \hbox{\char03}\hidewidth}}\fi}
 19. \def\o#1{\setbox\z@\hbox{#1}\dimen\z@=\dp\z@\ifdim\ht\z@=1ex\accent10 #1% ogonek
 20.   \else{\ooalign{\unhbox\z@\crcr\hidewidth\lower0.9\dimen\z@ %
 21.         \hbox{\char10}\hidewidth}}\fi}
 22. %
 23. \'A \u B \v C \^D \"E \.F \`H \H I \=J \r K \~L \d M \b N \c O \o P \o A --- 
 24. \'a \u b \v c \^d \"e \.f \`g \H h \=k \r l \~m \d n \b o \c p \o q \o a

Comments:

• The ``9y'' encoding is a custom encoding I made for Caslon 224.
• Lines 2-15 are a direct modification of the Plain TeX macros.
• Lines 16-18 modify Plain TeX's \c macro by adding the \lower command on line 17.
• Caslon 224 has an ogonek accent which is not in the Computer Modern fonts. The definition of \o on lines 19-21 is a straightforward modification of \c.
• Lines 23-24 use each accent to typeset a mixture of upper and lower case letters.

See Also: char.

Description:

THIS value is added to the demerits calculated for a line, when TeX is breaking a paragraph into lines, if the line and the previous one are visually incompatible [98]. During the break-up, TeX classifies each line as: tight, decent, loose, and very loose. Also, it says two consecutive lines are visually incompatible if their classifications are not adjacent (i.e., a tight line followed by a loose one, or a decent line followed by a very loose one) [97].

     \gdef\tstory{There are cries, sobs, confusion among the people,
     and at that moment the cardinal himself, the Grand Inquisitor, passes
     by the cathedral. He is an old man, almost ninety, tall and erect,
     with a withered face and sunken eyes, in which there is still a gleam
     of light. He is not dressed in his brilliant cardinal's robes, as he
     was the day before, when he was burning the enemies of the Roman
     Church~\char144\kern2em\hfill Fyodor Dostoyevsky, {\it The Brothers
     Karamazov}}
     \hsize=2.5in% The \baselineskip reference page holds the definition of \tstory
     \setbox0=\vbox{\adjdemerits=0
     \doublehyphendemerits=100000
     \finalhyphendemerits=900000
     \tstory\par}
     \setbox1=\vbox{\adjdemerits=1000000
     \doublehyphendemerits=100000
     \finalhyphendemerits=900000
     \tstory\par}
     \hbox{\box0\kern0.25in\box1}

Comments:

• This example continues the example on the \finalhyphendemerits reference page.
• The original paragraph (shown on the \doublehyphendemerits page) had several hyphen-related problems.
• Those problems are cleared up on the left paragraph. However, it still uses \adjdemerits=0.
• The paragraph on the right uses a large value for \adjdemerits and is slightly different from the left paragraph.

See Also: doublehyphendemerits, finalhyphendemerits.

Description:

THIS command is used to increase or decrease one of the <numeric variables>. The <quantity> should have a type compatible with the variable as shown in the following:


\advance<integer variable> by <number>

\advance<dimen variable> by <dimen>

\advance<glue variable> by <glue>

\advance<muglue variable> by <muglue>

An <integer variable> is: a) an explicit \count register (e.g., \count17); b) a control sequence specified using \countdef, \newcount, or something equivalent; or c) an integer parameter such as \tolerance or \widowpenalty. The dimen, glue, and muglue variables are similar. The ``by'' is optional [118 and 276].

  1. \def\tpagenumber#1% page number.
  2. {%
  3.   \bgroup
  4.      \count1=#1
  5.      \ifnum\count1 > 0
  6.           \edef\tpageA{\the\count1}%
  7.           \advance\count1 by 1
  8.           \def\tpageB{\the\count1}%
  9.      \else
 10.           \count1=-\count1% This makes the page > 0.
 11.           \edef\tpageA{\romannumeral\count1}%
 12.           \advance\count1 by 1
 13.           \def\tpageB{\romannumeral\count1}%
 14.      \fi
 15.      The current page number is: \tpageA. The next page number is: \tpageB.\par
 16.   \egroup
 17. }
 18. \tpagenumber{127}
 19. \tpagenumber{-15}

Comments:

• This example is not quite complete since the page number is not stored in a useful place.
• However, it shows one way to increase or decrease a counter and to typeset either the digits in a number or the letters in the roman numeral equivalent of the number.
• If \def is used instead of \edef on lines 6 and 11, the \tpageA macro will use the same value for \count1 that \tpageB uses.

See Also: multiply, divide, count, dimen, skip, muskip.

For Related Examples, see: day, ifnum, time

Description:

THIS command saves <token> (without expansion [215]) and inserts it back into the input after the next assignment. If two \afterassignment commands occur before an assignment, only the second token is saved. If the next assignment is `\setboxn=\hbox{...}', the saved token is inserted immediately after the `{' just before tokens from \everyhbox. Assignments made by \setbox with \vbox or \vtop work the same way [279].

  1. \def\ttrA#1{[#1]}
  2. \def\ttrB#1%
  3. {%
  4.      \setbox0=\hbox{{\selffamily AD [}}%
  5.      \setbox1=\hbox{{\selffamily AD ]}}%
  6.      \raise0.45\dp0\box0 %
  7.      {\selffamily AE  #1}%
  8.      \raise0.45\dp1\box1 %
  9. }%
 10. \def\trtC#1{\def\tpage{#1}\afterassignment\ttrD\let\next=}
 11. \def\ttrD
 12. {%
 13.      \advance\tnotenumber by 1%
 14.      \unskip
 15.      \if\next.% the workign version has other special cases: e.g., `)' `,'.
 16.           .\superscript{\the\tnotenumber}%
 17.      \else
 18.           \R1\superscript{\the\tnotenumber}\ \next
 19.      \fi
 20.      \global\setbox\tnotebox=\hbox
 21.      {%
 22.           \noindent
 23.           \unhbox\tnotebox
 24.           \superscript{\the\tnotenumber}%
 25.           \R1\tpage.\quad
 26.      }%
 27. }
 28. \def\texam#1{This example \tr{151} illustrates work-in-progress
 29. \tr{176--177}. Originally, all page references \tr{213 and 217} to
 30. {\it The TeXBook} \tr{399} in the reference pages appeared directly
 31. (e.g., |[151]|) \tr{255}. --- #1.}
 32. \let\tr=\ttrA
 33. \setbox0=\vbox{\texam1}
 34. \unvbox0\vskip1\baselineskip
 35. \let\tr=\ttrB
 36. \setbox0=\vbox{\texam2}
 37. \unvbox0\vskip1\baselineskip
 38. \let\tr=\trtC
 39. \tnotenumber=0
 40. \setbox\tnotebox=\hbox{}
 41. \setbox0=\vbox{\texam3}
 42. \unvbox0\vskip0.5\baselineskip\unhbox\tnotebox

Comments:

• This is a long example, but it illustrates the power of \afterassignment, \let, \if and TeX in a real situation. The same source (lines 28-31) is made to appear in three distinct ways by changing a formatting routine.
• The basic question is ``how should page references to The TeXbook appear in the Description area on a reference page?'' The example shows three possibilities.
• The first paragraph is made by lines 1 and 28-34. In my opinion the brackets and page numbers in it are much too large.
• The second paragraph is made by lines 2-9, 28-31, and 35-37. I used trial-and-error before deciding on the values used on lines 4-8. Most of the reference pages were completed using this style before I decided there were enough entries like references 2, 3, and 5 to warrent giving the matter more attention.
• The third paragraph is the result of that attention. It provides complete documentation in an unobstructive way (at least in my opinion).
• The one wrinkle is that footnotes should come after periods. After some more thought, I realized TeX could swap the period and the footnote on an as needed basis. That is the real point of this example.
• The third paragraph is made by lines 10-31 and 38-42. The macro \ttrC forms the link between the document and \ttrD which does the typesetting. First, \ttrC saves the reference in a temporary macro. Next, it says: ``read the first token after this macro and put it in \next; then do \ttrD.'' The next token will be a period (for notes 2 and 5) or the character following `}\]' for the other notes (e.g., an `i' for note 1).
• Line 14 removes the glue made by the space which preceeds each \ttr{} in the paragraph.
• Line 15 compares \next and `.'. If they are the same, TeX typesets a period and the correct footnote number. Otherwise, it inserts a small amount of glue (the \R1) and typesets the correct footnote number, a control space, and \next (the next token).
• Lines 20-26 are a quick and dirty way to build the line of references one at a time.

See Also: aftergroup, expandafter, futurelet, noexpand.

Description:

THIS command saves <token> on TeX's save stack (without expansion [215]) and inserts it back into the input when the current group ends. If several \aftergroup commands occur in a group, each of the saved tokens is inserted in the same order the commands occurred [279].

  1. \def\ta{Hello}
  2. \def\tb{ World. }
  3. \def\tc{How are you?}
  4. \begingroup
  5. \aftergroup\ta
  6. \aftergroup\tb
  7. \aftergroup\tc
  8. \let\ua=\tc
  9. \let\ub=\tb
 10. \let\uc=\ta
 11. \gdef\ta{I am}
 12. \gdef\tb{ fine. }
 13. \ua\ub\uc---\par
 14. \endgroup
 15. ---

Comments:

• Lines 5-7 setup three aftergroup tokens, and lines 11-12 change two of the tokens.
• When the group ends, all three tokens appear with the correct values.

See Also: afterassignment, expandafter, futurelet, noexpand.

Description:

THIS command places one subformula over a second subformula, only it doesn't put a line between the two subformulas as \over does [143]. The command is an abbreviation for `\atopwithdelims..' [152].

  1. $n \atop k = n! / k!(n-k)!,$
  2. \hbox{this doesn't work. It needs a pair of braces.}\par\vskip0.5\baselineskip
  3. ${n \atop k} = n! / k!(n-k)!$, this looks okay, but it is missing ()'s.
  4. $$n \atop k = n! / k!(n-k)!,\quad
  5. \hbox{this doesn't work. It needs a pair of braces.}$$\par
  6. $${n \atop k} = {n! \over k!(n-k)!} = n! / k!(n-k)!,\quad
  7. \hbox{this looks okay, but it is missing ()'s.}$$

Comments:

• TeX's generalized fraction commands often require braces. Lines 1 and 4-5 show what happens if a required pair is omitted.
• Lines 3 and 6-7 show the corrected material.
• Several of the fractions shown above should be surrounded by parenthesis. They are provided by \atopwithdelims.

See Also: over, above, atopwithdelims.

Description:

THIS command is a generalized form of \atop. It builds the same fraction \atop does (without a line). Then, it puts <delim1> before the fraction and <delim2> after it. Both delimiters are scaled appropriately. If a `.' is used for either <delim>, TeX ignores the corresponding delimiter. In fact, \atop is an abbreviation for `\atopwithdelims..'. Also, \atopwithdelims is derived from \abovewithdelims [152].

  1. \def\tchoose{\atopwithdelims()}
  2. \def\tc#1#2{{#1\tchoose #2}}
  3. $\tc nk = {n \tchoose n-k} = n! / k!(n-k)!$\quad now have needed parenthesis.
  4. $$\displaylines{\hbox{Also,}\qquad{n \tchoose k} = {n! \over k!(n-k)!}= 
  5. n!/k!(n-k!).\hfill\cr}$$
  6. {\bf Theorem 75}.\quad If $p$ is a prime, then 
  7. $$\tc p1,\ \tc p2,\ldots,\ \tc p{p-1}$$ 
  8. are divisible by $p$.

Comments:

• Lines 1-2 provide an alternate to Plain TeX's \choose macro.
• The utility of \tc becomes apparent in the display made by line 7.

See Also: atop, abovewithdelims.

Description:

THIS is an internal quantity set after each box is made. It is between 0 and 10,000 unless the box is overfull. Then it is 1,000,000 [229]. If r is the glue set ratio [77], \badness is approximately min(100r³,10000) [97].

     \def\badboxes#1#2%
     {%
          \overfullrule=0.25pt
          \setbox0=\hbox spread#2{#1}%
          \count12=\the\badness
          \setbox1=\hbox to 2.5in{\box0\hfil\the\count12}%
          \box1
     }
     \badboxes{The badness of this line is: }{-1em}
     \badboxes{The badness of this line is: }{-0.5em}
     \badboxes{The badness of this line is: }{-0.3em}
     \badboxes{The badness of this line is: }{0em}
     \badboxes{The badness of this line is: }{1em}
     \badboxes{The badness of this line is: }{2em}
     \badboxes{The badness of this line is: }{3em}

Comments:

• The first line with a badness of 1,000,000 has a very thin rule after `is: '. It is the rule placed at the end of overfull boxes. Normally, it is 5 points wide, but the \overfullrule command in \badboxes changed the width to 0.25 points.
• Two of the lines have the same \badness. However, one line is too tight and the other is too loose. So, \badness, by itself, cannot distinguish between tight and loose lines.

See Also: hbadness, vbadness, pretolerance, tolerance, emergencystretch.

Description:

THIS parameter gives the desired glue which should separate boxes in a vertical list [78]. In actual practice TeX uses three other primitives to compute the interline glue placed between boxes: \prevdepth = p, \lineskip, and \lineskiplimit = l. Let the box about to be added have height h, and let \baselineskip = {b plus x minus y}. If p <= -1000pt, no interline glue is added. Otherwise, if b >= p + h + l, the glue `(b-p-h) plus x minus y' is added just before the new box. Otherwise, the \lineskip glue is placed just before the new box. After the new box is added, \prevdepth is set to its depth [80]. Text set using 10-point type often uses `\baselineskip = 12pt'. The same ratio (1.2) works well with point sizes from 7-14 points [nr]. To suppress interline glue between two boxes put `\prevdepth=-1000pt' just before the second one. To suppress interline glue between a number of boxes set `\baselineskip=-\maxdimen', `\lineskiplimit=\maxdimen', and `\lineksip=0pt' before the first box [352].

     \parindent=2pc \hsize=4.75in% Point size in all examples is 10pt.
     \baselineskip=11.5pt\tstory\par
     \baselineskip=12pt\tstory\par
     \baselineskip=12.5pt\tstory\par

Comments:

• These paragraphs show the effect of a small change in the space put between lines of text. The results are even more striking if several pages are typeset each with its own \baselineskip value.

See Also: lineskip, lineskiplimit, prevdepth.

Description:

NORMALLY, when TeX discovers an error in a file, it stops processing, displays a message on the terminal, and waits for instructions for what to do next. One option at this point is to press Q (run quietly). That tells TeX to continue processing without stopping if further errors arise and to suppress further output to the terminal [31]. If TeX encounters \batchmode in a file, then it begins to process the file just as if Q was entered in response to an error. This is a global change which continues until the end of the file or until one of the other `..mode' commands is encountered [32]. Certain errors cause TeX to stop processing prematurely even in \batchmode (e.g., a missing file on an \input statement) [299].

  1. \batchmode
  2.   .
  3.   .
  4.   .
  5. \errorstopmode

Comments:

• If TeX encounters errors on lines 2-4, it writes an error message to the log file, but it won't display a message on the terminal or stop for user input.
• After line 5 TeX resumes normal error processing.

See Also: errorstopmode, nonstopmode, scrollmode.

Description:

THIS command begins a new level of grouping. That group must be terminated by \endgroup not by `}'. The mode doesn't change [279]. Groups begun by `{' and \begingroup must be properly nested; `{\begingroup}\endgroup' generates an error [21].

  1. \catcode`\?=13
  2. \def?{\begingroup\ccflip\obeyspaces\getword}%
  3. \def\getword#1?{#1\endgroup}%
  4. Hello ?\World? How are you\char63\kern2.5pt?\char63? See you later.

Comments:

• Line 1 makes `?' an active character.
• Line 2 defines `?'. The definition begins a new group; changes the catcodes of the characters used to control TeX ; and calls \getword.
• Line 3 shows \getword has a single parameter which is ended by a `?'. Thus, everything between the two `?'s gets typeset verbatim. The \endgroup on line 3 matches the \begingroup on line 2. When \getword and the group ends, each of the catcodes changed on line 2 reverts to its normal value.
• Line 4 holds some sample text. Note, the ASCII value of a `?' is 63.

See Also: endgroup.

Description:

AFTER TeX processes a math display, it adds the following items to the current vertical list: a penalty, glue, the actual display, glue, and a penalty. The two glue items are either the above and below displayskips or the above and below displayshortskips. The latter are used if: a) the display occurs at the start of a paragraph; b) the display follows another display; or c) the display occurs in the middle of a paragraph but does not have a left equation number and does not overlap the preceeding line of text [189].

  1. \def\tfns#1#2#3#4#5#6% 4 skips, post-example skip, L/R equation # code.
  2. {%
  3.      \abovedisplayskip=#1
  4.      \belowdisplayskip=#2
  5.      \abovedisplayshortskip=#3
  6.      \belowdisplayshortskip=#4
  7.      \hfill The Fibonacci numbers are defined\break 
  8.      by: $F_0=1$, $F_1=1$, and
  9.      $$F_n = F_{n-1} + F_{n-2} \hbox{, for } n \ge 2.
 10.      \def\temptyF{#6}
 11.      \ifx\temptyF\empty
 12.           \leqno(13.12)$$
 13.      \else
 14.           \eqno(13.12)$$
 15.      \fi
 16.      A related sequence $\{v_n\}$ is: 2, 1, 3, 4, 7, 11, 18, \char144\ It
 17.      satisfies $v_n$ = $x^n + y^n$, where $x = (1 + \sqrt 5)/ 2$ 
 18.      and $y = (1 - \sqrt 5)/ 2$.
 19.      \vskip#5
 20. }
 21. \tfns{12pt}{12pt}{3pt}{3pt}{1\baselineskip}{}\par
 22. \tfns{12pt}{12pt}{3pt}{3pt}{0pt}{-}

Comments:

• This example is slightly more complicated than the other display skip examples. It uses lines 10-15 to insert an equation number either on the left or on the right of the display.
• Neither of the displays overlaps the previous line of text. However, the display made by line 21 has a left equation number and uses the regular (non-short) glue. The display made by line 22 has a right equation number and uses short glue.

See Also: abovedisplayshortskip, belowdisplayskip.

Description:

AFTER TeX processes a math display, it adds the following items to the current vertical list: a penalty, glue, the actual display, glue, and a penalty. The two glue items are either the above and below displayskips or the above and below displayshortskips. The former are used if: a) the display occurs in the middle of a paragraph and the display overlaps the preceeding line of text; or b) the display has a left equation number [189]. Also, the former are used before and after an alignment display [190].

     \def\tfn#1#2#3#4#5% above & below regular & short skips + post-example skip
     {%
        \abovedisplayskip=#1
        \belowdisplayskip=#2
        \abovedisplayshortskip=#3
        \belowdisplayshortskip=#4
        The Fibonacci numbers are defined by: $F_0=1$, $F_1=1$, and
        $$F_n = F_{n-1} + F_{n-2} \hbox{, for } n \ge 2.\leqno(13.11)$$
        A related sequence $\{v_n\}$ is: 2, 1, 3, 4, 7, 11, 18, \char144\ It
        satisfies $v_n$ = $x^n + y^n$, where $x = (1 + \sqrt 5)/ 2$ 
        and $y = (1 - \sqrt 5)/ 2$.
        \vskip#5
     }
     \tfn{3pt}{3pt}{6pt}{6pt}{1\baselineskip}
     \tfn{12pt}{12pt}{6pt}{6pt}{0pt}

Comments:

• This example is very similar to the one for \abovedisplayskip. It uses the regular (non-short) display skips.
• The first example shows what Plain TeX makes if the glue shrinks its maximum amount, and the second example shows what happens if Plain TeX's glue neither stretches nor shrinks.

See Also: abovedisplayskip, belowdisplayshortskip.

Description:

JUST as TeX does not randomly hyphenate words, it does not randomly break math expressions between lines in a paragraph. TeX only considers a break after a relation (e.g., = or <) or a binary operation (e.g., + or ×) [173]. TeX inserts a penalty after a binary operation [174] which is not enclosed in braces and is not part of a generalized fraction [173]. The actual value of the penalty added after a binary operation in an expression is the value of \binoppenalty at the end of the expression [101]. However, the penalty is not inserted if: a) the binary operation is the final item in the entire list; b) the value of the penalty is 10,000 or more; or c) the next item in the list is a penalty item [446-447].

See Also: relpenalty.

Description:

THIS command holds the mark text that was most recently encountered on the page that was just boxed. It is null (i.e., empty) before the first page. It is equal to the previous \botmark, when a page contains no marks [258]. The value of \botmark is global (i.e., it is not affected by grouping) [259].

  1. % These lines are from the macro that makes the reference page
  2. % for a control sequence.
  3. % We have: \def\CScs{ControlSequence}, e.g., \def\CScs{botmark}
  4. % \box0 holds everything through the Synopsis line.
  5.      \mark{\currentcs\noexpand\else\CScs}
  6. % Do \eject unless following \box0 will fit on current page.
  7. % The \ifdim reference page shows this test.
  8.      \unvbox0%
  9.      \mark{\CScs\noexpand\else\CScs}
 10. % Typeset the rest of the page.
 11.      \let\currentcs=\CScs

Comments:

• It is tricky to get the correct guide word on the header line of these reference pages. Two things are required [260-261].
• First, lines 5 and 9 put a special mark before and after the start of each control sequence. Each mark contains two options separated by `\noexpand\else'.
• Second, the header routine for recto pages uses `\iftrue\botmark\fi' to select that page's guide word.
• The `if' statement expands to \currentcs unless the test referred to on lines 6-7 shows there is room for the start of the next command on the current page. If there is, the mark on line 9 appears on the current page, and the `if' statement expands to \CScs.

See Also: mark, firstmark, topmark.

Description:

THERE are 256 box registers: \box0 to \box255. A \box command adds the contents of the box to the current list, and it empties the box. It often is better to use \unhbox or \unvbox since they remove the outermost layer of glue in the box. That allows the box's contents to mesh with the other items in the current list. The \copy command is similar to \box except \copy does not empty the box [120]. Plain TeX has a \newbox command which reserves a box for a special purpose [121]. Each insertion class (\insert) reserves a box for its use [122]. Also, \box255 holds the current page which is used by the \output routine [253].

  1. \setbox0=\hbox{\selffamily AB ``Hello, World.''}
  2. \box0\par
  3. \setbox1=\hbox{``You were in a box, but now you are not!''}
  4. \box1\box0

Comments:

• The \box0 command on line 4 does nothing because \box0 was emptied on line 2. See the \copy page for comparison.

See Also: copy, unhbox, unvbox, unhcopy, unvcopy.