| 1 |
|
| 2 |
\input minim-doc.sty |
| 3 |
|
| 4 |
\manual{minim-math} |
| 5 |
|
| 6 |
\ifchapter \chapter Mathematics |
| 7 |
|
| 8 |
\else \startmetadata |
| 9 |
author {Esger Renkema} |
| 10 |
title {minim-math} |
| 11 |
date {2025-02-12} |
| 12 |
version {2025/1.5} |
| 13 |
keywords {LuaTeX; Plain TeX; Unicode mathematics} |
| 14 |
stopmetadata |
| 15 |
\maketitle \fi |
| 16 |
|
| 17 |
This package gives a simple and higly-configurable way of using unicode and |
| 18 |
OpenType mathematics with plain LuaTeX, making use of most of the latter |
| 19 |
engine’s new capabilities in mathematical typesetting. Also included are proper |
| 20 |
settings and definitions for nearly all unicode mathematical characters, as |
| 21 |
well as a few shorthands and helper macros that seemed useful additions. |
| 22 |
|
| 23 |
Load the package by saying ⟦\input minim-math.tex⟧; this will set up luatex for |
| 24 |
using opentype mathematical fonts and unicode math input. |
| 25 |
It will not, however, select mathematical fonts for you. That you will have to |
| 26 |
do for yourself; see below for instructions. |
| 27 |
|
| 28 |
|
| 29 |
\section Experimental LaTeX support |
| 30 |
|
| 31 |
As of version 1.5, an experimental latex package is available as a faster |
| 32 |
alternative to unicode-math. |
| 33 |
The unicode-math user interface is not replicated though: it adds nothing to |
| 34 |
the plain tex version but the ⟦\setfontfam⟧ macro for loading a math font. |
| 35 |
Use it like this: |
| 36 |
|
| 37 |
{\leftskip2em |
| 38 |
⟦\usepackage {minim-mp} |
| 39 |
% additional fonts can be loaded with \setfontfam |
| 40 |
\textfam = 1 % for the \text command; see below |
| 41 |
\setfontfam \textfam {Latin Modern Roman} |
| 42 |
% the next line is also the default |
| 43 |
\setfontfam 0 {Latin Modern Math} ⟧ |
| 44 |
\par} |
| 45 |
|
| 46 |
Bugs are likely many at this point; reports are welcome. |
| 47 |
|
| 48 |
|
| 49 |
\section Styles and alphabets |
| 50 |
|
| 51 |
For some (mostly alphabetical) characters, multiple variants are available, |
| 52 |
e.g. $\mit A \mup A \mbf A \mbfit A \mbb A \frak A \bffrak A \scr A \bfscr A$. |
| 53 |
You can (locally) override the default style of these with |
| 54 |
⟦*\mathstyle {style}⟧ (equivalent to the old ⟦\bf⟧, ⟦\rm⟧ etc.) |
| 55 |
or with one of the shorthands that apply the style to their argument only: |
| 56 |
\smallskip\smallskip\automarktable \halign{\qquad |
| 57 |
#\unskip\quad\hfil &#\unskip\quad\hfil &#\unskip\hfil\cr |
| 58 |
Shorthand\quad\strut & Value of ⟦style⟧\quad\strut & Result \cr |
| 59 |
⟦*\mup⟧ & ⟦up⟧/⟦rm⟧ & $\mup {ABC}$ \cr |
| 60 |
⟦*\mit⟧ & ⟦it⟧ & $\mit {ABC}$ \cr |
| 61 |
⟦*\mbf⟧ & ⟦bf⟧ & $\mbf {ABC}$ \cr |
| 62 |
⟦*\mbfit⟧ & ⟦bfit⟧ & $\mbfit {ABC}$ \cr |
| 63 |
⟦*\mbb⟧ & ⟦bb⟧ & $\mbb {ABC}$ \cr |
| 64 |
⟦*\frak⟧ & ⟦frak⟧ & $\frak {ABC}$ \cr |
| 65 |
⟦*\bffrak⟧ & ⟦bffrak⟧ & $\bffrak {ABC}$ \cr |
| 66 |
⟦*\scr⟧ & ⟦cal⟧/⟦scr⟧ & $\scr {ABC}$ \cr |
| 67 |
⟦*\bfscr⟧ & ⟦bfscr⟧ & $\bfscr {ABC}$ \cr |
| 68 |
} |
| 69 |
Styles without shorthand are ⟦sans⟧/⟦sf⟧, ⟦sfit⟧, ⟦sfbf⟧, ⟦sfbfit⟧, ⟦tt⟧/⟦mono⟧ |
| 70 |
and finally the special value ⟦*clear⟧ for using the default style. |
| 71 |
You can use the shorthands directly in sub- and superscripts: ⟦v^\scr F⟧ will |
| 72 |
result in $v^\scr F$\kern-4pt. |
| 73 |
|
| 74 |
While math families are not used anymore for switching between styles (see |
| 75 |
below), you still can use ⟦\fam⟧ with the values 0, 1, 2, 4, 5, 6 or~7 for |
| 76 |
doing so. This means that plain tex’s ⟦\rm⟧, ⟦\it⟧, ⟦\cal⟧, ⟦\sl⟧, ⟦\bf⟧ and |
| 77 |
⟦\tt⟧ can still be used (at least in math mode). |
| 78 |
|
| 79 |
Please note that ⟦*\mup⟧ is not the right choice for upright multiletter |
| 80 |
symbols or operators: you should use ⟦*\mord⟧ or ⟦*\mop⟧ instead (see near the |
| 81 |
end of this chapter). |
| 82 |
For nonmathematical text, you should use ⟦*\text⟧ instead of ⟦\mup⟧. |
| 83 |
|
| 84 |
The default properties of characters can be set with one of the following three |
| 85 |
commands: |
| 86 |
\startlist |
| 87 |
\tagattribute List ListNumbering /Unordered |
| 88 |
\item\ignore. ⟦*\mathmap {character list} {style}⟧ |
| 89 |
\item\ignore. ⟦*\mathclass {character list} {class}⟧ |
| 90 |
\item\ignore. ⟦*\mathfam {character list} nr⟧ |
| 91 |
\stoplist |
| 92 |
There, ⟦style⟧ is one of the above and ⟦class⟧ is the name of a class as below. |
| 93 |
Finally, the ⟦character list⟧ should be a comma-separated list with elements of |
| 94 |
one of the following forms: |
| 95 |
\startlist |
| 96 |
\tagattribute List ListNumbering /Decimal |
| 97 |
{\def\⟦#1{⟦[\uppercase{#1}#1]\kern-1pt} |
| 98 |
\item1. a list of characters, like ⟦abc⟧ or ⟦\partial⟧ or $ℝ$; |
| 99 |
\item2. a character range, like ⟦`A-`Z⟧, ⟦65-90⟧ or ⟦"41-"5A⟧; |
| 100 |
\item3. one of the alphabets \⟦latin⟧, \⟦greek⟧, or ⟦digits⟧; |
| 101 |
\item4. one of the style groups ⟦bold⟧, ⟦boldgreek⟧, ⟦sans⟧, ⟦sansgreek⟧, |
| 102 |
⟦mono⟧, ⟦blackboard⟧, ⟦fraktur⟧ or ⟦script⟧; |
| 103 |
\item5. the name of a math class: |
| 104 |
⟦ord⟧, ⟦op⟧, ⟦bin⟧, ⟦rel⟧, ⟦fence⟧, ⟦open⟧, ⟦close⟧ or ⟦punct⟧. |
| 105 |
}\stoplist |
| 106 |
|
| 107 |
Note that unicode is somewhat irregular in its encoding of mathematical |
| 108 |
letters; this is taken into account when using ranges as under~(2) above. Thus, |
| 109 |
⟦`\mscra-`\mscrz⟧ really gives you all lowercase script characters, despite |
| 110 |
e.g.~$ℯ$ being well outside that range. |
| 111 |
|
| 112 |
The default style settings are ⟦\mathmap {latin, greek, Latin}{it}⟧. |
| 113 |
Since the math family setting is not used anymore for selecting different |
| 114 |
styles, the default family of every symbol is zero. |
| 115 |
Instead, you can use ⟦\mathfam⟧ for mixing fonts (see below). |
| 116 |
The ⟦class⟧ option to ⟦\mathclass⟧ should be one of the names under~5. |
| 117 |
|
| 118 |
|
| 119 |
\section Character variants |
| 120 |
|
| 121 |
You can change the default appearance of several greek characters with |
| 122 |
⟦*\use∥-math∥-variant {chars}⟧ or ⟦*\usemathdefault {chars}⟧, where ⟦chars⟧ is |
| 123 |
a list of normal greek characters. |
| 124 |
As in unicode but against tex’s tradition, the variants are |
| 125 |
$\usemathvariant{εθΘκπρφ}εθΘκπρφ$ and the defaults $εθΘκπρφ$. |
| 126 |
The macros ⟦*\varepsilon⟧ etc. have been updated to reflect the unicode |
| 127 |
variants. |
| 128 |
|
| 129 |
The appearance of root symbols can be set with ⟦*\closedroots⟧ ($\closedroots |
| 130 |
\smash{√2}$) and ⟦*\normalroots⟧ ($\smash{√2}$, the default). |
| 131 |
|
| 132 |
Say ⟦*\unicodedots⟧ to use the unicode dots characters ($\unicodedots…⋮⋰⋱$) and |
| 133 |
⟦*\traditionaldots⟧ to construct these characters from periods ($\smash{…⋮⋰⋱}$, |
| 134 |
the default). |
| 135 |
Both settings affect the meaning of both the actual characters and the ⟦\xdots⟧ |
| 136 |
macros ($\mathstyle{mono}x∈\{l,v,c,a,d\}$). |
| 137 |
Unlike in traditional plain tex, the traditional dots are available in script |
| 138 |
sizes, too. |
| 139 |
|
| 140 |
Say ⟦*\decimalcomma⟧ and have commas appear as $\decimalcomma1,2$ instead |
| 141 |
of~$1,2$ (⟦*\nodecimalcomma⟧ restores the default). |
| 142 |
Say ⟦*\smartdecimalcomma⟧ for a comma that only acts as punctuation when not |
| 143 |
immediately followed by a digit. |
| 144 |
The explicit ⟦*\comma⟧, like ⟦*\colon⟧, will always be punctuation. |
| 145 |
|
| 146 |
The behaviour of limits on integral signs can be set by redefining |
| 147 |
⟦*\intlimits⟧ (the default is ⟦\let \intlimits = \nolimits⟧). |
| 148 |
|
| 149 |
If you want to change the meaning (inside math mode) altogether for a single |
| 150 |
character, you can use the commands ⟦*\mathdef⟧ and ⟦*\mathlet⟧. |
| 151 |
For example, by default, you can use the letter ħ for the reduced planck |
| 152 |
constant~$ħ$; |
| 153 |
this has been made so with ⟦*\mathdef ħ {\hbar}⟧ (you could also have said |
| 154 |
⟦\mathlet ħ \hbar⟧). |
| 155 |
|
| 156 |
|
| 157 |
\section Setting up fonts |
| 158 |
|
| 159 |
The minimum you need do to set up a mathematical font is this: |
| 160 |
|
| 161 |
⟦*\font\tenmath |
| 162 |
{Latin Modern Math:mode=base;script=math;ssty=0;} at 10pt |
| 163 |
\font\tenmaths |
| 164 |
{Latin Modern Math:mode=base;script=math;ssty=1;} at 7pt |
| 165 |
\font\tenmathss |
| 166 |
{Latin Modern Math:mode=base;script=math;ssty=2;} at 5pt |
| 167 |
\textfont 0 = \tenmath |
| 168 |
\scriptfont 0 = \tenmaths |
| 169 |
\scriptscriptfont 0 = \tenmathss⟧ |
| 170 |
|
| 171 |
Note that you only have to set up the font for a single family: opentype |
| 172 |
mathematical fonts typically contain all necessary variants of all mathematical |
| 173 |
characters. |
| 174 |
Therefore, the ⟦*\fam⟧ setting has been made a no-op (use ⟦*\setfam⟧ if you |
| 175 |
really need the old primitive) and the default family of all symbols has been |
| 176 |
set to zero. |
| 177 |
|
| 178 |
As mentioned above, you can still change the family number of specific |
| 179 |
characters and this allows you to mix mathematical fonts. For instance, if you |
| 180 |
dislike the current blackboard bold characters, just assign a second font to |
| 181 |
family |
| 182 |
1 and say ⟦\mathfam {blackboard} 1⟧. Less useful are the parameters |
| 183 |
⟦*\accentfam⟧, ⟦*\radicalfam⟧ and ⟦*\extensiblefam⟧ that control the family of |
| 184 |
all accents, radicals and extensibles. |
| 185 |
|
| 186 |
Do note that various spacing constants are set according to the \emph{last} |
| 187 |
math family that is assigned to. Therefore, you should assign your main math |
| 188 |
font to a family after all others. |
| 189 |
|
| 190 |
You can use ⟦*\text⟧ for adding nonmathematical text to your equations. |
| 191 |
It will behave well in sub- and superscripts: |
| 192 |
⟦\text{word}^\text{word}⟧ gives $\smash{\text{word}^\text{word}}$. |
| 193 |
By default, the font used is the normal mathematical font. You can change this |
| 194 |
by setting the ⟦*\textfam⟧ parameter to some nonzero value and assigning |
| 195 |
a different font to that family. |
| 196 |
You probably want this, since most commonly-used mathematical fonts do not |
| 197 |
include a normal kerning table. |
| 198 |
|
| 199 |
\section Shorthands and additions |
| 200 |
|
| 201 |
All the usual arrows can be made extensible by prefacing them with an ⟦x⟧, |
| 202 |
including ⟦\xmapsto⟧ and ⟦\xmapsfrom⟧. Alternatively, you can use ⟦\→⟧ etc. as |
| 203 |
shorthands. Additionally, you can use the following: |
| 204 |
|
| 205 |
{\smallskip\smallskip\automarktable \halign{\qquad |
| 206 |
\vrule width0pt depth 9pt#\unskip\hfil |
| 207 |
&&\qquad#\unskip\hfil\cr |
| 208 |
Shorthand & Result \cr |
| 209 |
⟦\bra x, \ket y⟧ & $\bra x, \ket y$ \cr |
| 210 |
⟦\braket x y⟧ & $\braket x y$ \cr |
| 211 |
⟦\norm x, \Norm x⟧ & $\norm x, \Norm x$ \cr |
| 212 |
⟦\floor x, \ceil x⟧ & $\floor x, \ceil x$ \cr |
| 213 |
⟦x \stackrel ?= y⟧ & $x \stackrel ?= y$ \cr |
| 214 |
⟦x \stackbin a+ y⟧ & $x \stackbin a+ y$ \cr |
| 215 |
⟦f\inv⟧ & $f\inv$ \quad (cf. $f^{-1}$) \cr |
| 216 |
⟦a \xrightarrow[down]{up} b⟧ & $a \xrightarrow[down]{up} b$ \cr |
| 217 |
⟦a \xeq[down]{up} b⟧ & $a \xeq[down]{up} b$ \cr |
| 218 |
⟦\frac12, \tfrac12, \dfrac12⟧& $\frac12, \tfrac12, \dfrac12$\cr |
| 219 |
}\smallskip} |
| 220 |
|
| 221 |
Also new are the operators |
| 222 |
⟦\Tr⟧, ⟦\tr⟧, ⟦\Span⟧, ⟦\GL⟧, ⟦\SL⟧, ⟦\SU⟧, ⟦\U⟧, ⟦\SO⟧, ⟦\O⟧, |
| 223 |
⟦\Sp⟧, ⟦\im⟧, ⟦\End⟧, ⟦\Aut⟧, ⟦\Dom⟧ and ⟦\Codom⟧. |
| 224 |
You can define new operators with ⟦*\newmathop⟧ and ⟦*\newlargemathop⟧: |
| 225 |
⟦\newmathop{op}⟧ will define the new operator ⟦\op⟧. |
| 226 |
For single use of an upright symbol, operator or large operator you can use |
| 227 |
⟦*\mord⟧, ⟦*\mop⟧ and ⟦*\mlop⟧. |
| 228 |
The difference between ⟦\mord⟧ and ⟦\mup⟧ is that ⟦\mord⟧ also applies the |
| 229 |
correct symbol spacing. |
| 230 |
|
| 231 |
The accents ⟦\overbrace⟧, ⟦\underbracket⟧ etc. allow a label between square |
| 232 |
brackets: |
| 233 |
⟦$$\underbrace[=1]{(x^2+∥penalty0y^2)}$$⟧ gives |
| 234 |
$$\underbrace[=1]{(x^2 + y^2)}.$$ |
| 235 |
|
| 236 |
Finally, for plain tex only, the following (entirely optional) alternative to |
| 237 |
using dollar signs is provided, which also offers slight improvements in the |
| 238 |
spacing of displayed equations: |
| 239 |
\stopformulatagging\Ustartdisplaymath |
| 240 |
\halign{# \hfil&#\quad&#\hfil\cr |
| 241 |
&⟦\[ ... \]⟧&inline mathematics\cr |
| 242 |
⟦\display⟧&⟦\[ ... \]⟧&display mathematics\cr |
| 243 |
⟦\displaynr {nr}⟧&⟦\[ ... \]⟧&numbered display mathematics\cr |
| 244 |
⟦\displaynr⟧&⟦\[ ... \]⟧&automatically numbered display mathematics\cr |
| 245 |
}\Ustopdisplaymath\startformulatagging |
| 246 |
|
| 247 |
The automatic display numbering uses the count ⟦\equationnumber⟧ and the token |
| 248 |
list ⟦\setequationnumber⟧ internally. All displays created this way can be made |
| 249 |
left-aligned by saying ⟦\leftdisplaystrue⟧. In that case, the displays are |
| 250 |
offset from the margin by ⟦\leftdisplayindent⟧ (a muskip, default value~⟦0mu⟧). |
| 251 |
|
| 252 |
\section Best practices |
| 253 |
|
| 254 |
The following remarks on mathematical typesetting have no relation to the |
| 255 |
contents of this package; I have included them because I find them hard to |
| 256 |
remember. |
| 257 |
|
| 258 |
\startlist |
| 259 |
\tagattribute List ListNumbering /Decimal |
| 260 |
\item1. ⟦*\eqalign⟧ gives a vertically centered box and can occur many times in |
| 261 |
an equation, while ⟦*\eqalignno⟧ and ⟦*\leqalignno⟧ span whole lines (put the |
| 262 |
equation numbers in the third column). All assume the relation (or operator) |
| 263 |
appears at the right hand side of the ampsersand. |
| 264 |
\item2. The command ⟦*\displaylines⟧ can only have one column that spans the |
| 265 |
whole line (and you will have to add the equation number by hand). |
| 266 |
\item3. Further alignment commands are ⟦*\cases⟧, ⟦*\matrix⟧, ⟦*\pmatrix⟧ (with |
| 267 |
parentheses) and ⟦*\bordermatrix⟧ (includes labels for lines and columns). |
| 268 |
\item4. Finetuning alignments can be done with ⟦*\smash⟧, ⟦*\phantom⟧, |
| 269 |
⟦*\hphantom⟧ and ⟦*\vphantom⟧. |
| 270 |
\item5. Small matrices like $1\;2\choose3\;4$ can be made by misusing |
| 271 |
⟦*\choose⟧ or ⟦*\atop⟧. |
| 272 |
\item6. If you start a line with a binary operator, put a ⟦*{}⟧ before it: this |
| 273 |
way, tex recognises it as such. |
| 274 |
\item7. Thin spaces (⟦*\,⟧) should be inserted: before $\mup dx$, before units, |
| 275 |
after factorials and after ⟦*\dots⟧ if those are followed by a closing |
| 276 |
parenthesis. |
| 277 |
\item8. Whether the differential operator should be set upright or not is as of |
| 278 |
yet an open question in mathematics. |
| 279 |
\item9. You should prefer ⟦*\bigr⟧ and ⟦*\bigl⟧ etc. over ⟦\big⟧, ⟦\Big⟧, |
| 280 |
⟦\bigg⟧ and ⟦\Bigg⟧. |
| 281 |
\item10. An overview of mathematical symbols, with control sequences and their |
| 282 |
availability in different fonts, can be found in ⟦*unimath-symbols.pdf⟧, which |
| 283 |
is part of the unicode-math package. |
| 284 |
\stoplist |
| 285 |
|
| 286 |
\endmanual |
| 287 |
|
| 288 |
|