| 1 |
%% \CheckSum{145} |
| 2 |
% \iffalse |
| 3 |
% File: paramcalc.dtx |
| 4 |
% Copyright (C) 2026 Miguel V. S. Frasson ([email protected]) |
| 5 |
% |
| 6 |
% This package may be distributed under the terms of the LaTeX |
| 7 |
% Project Public License, as described in lppl.txt in the base |
| 8 |
% LaTeX distribution, either version 1.3 or (at your option) |
| 9 |
% any later version. |
| 10 |
% |
| 11 |
%<*driver> |
| 12 |
\documentclass{ltxdoc} |
| 13 |
\usepackage[T1]{fontenc} |
| 14 |
\usepackage[lining]{ebgaramond} |
| 15 |
\usepackage{ebgaramond-maths,textcomp} |
| 16 |
\usepackage[cmintegrals,cmbraces]{newtxmath} |
| 17 |
\let\epsilon\varepsilon |
| 18 |
\usepackage{xcolor,microtype,amsmath,amstext,tikz,centerlastline} |
| 19 |
\usepackage{paramcalc} |
| 20 |
% \usepackage[margin=1in]{geometry} |
| 21 |
\newcommand{\var}[1]{\texttt{\expandafter\string\csname #1\endcsname}} |
| 22 |
\RecordChanges |
| 23 |
\begin{document} |
| 24 |
\DocInput{paramcalc.dtx} |
| 25 |
\end{document} |
| 26 |
%</driver> |
| 27 |
% \fi |
| 28 |
% |
| 29 |
% %%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%% |
| 30 |
% |
| 31 |
% \StopEventually |
| 32 |
% |
| 33 |
% \changes{1.0}{2026/01/27}{Initial version} |
| 34 |
% |
| 35 |
% \title{\textsc{paramcalc}: compute parameters to fit condition} |
| 36 |
% \author{Miguel V.\ S.\ Frasson\\(\texttt{[email protected]})} |
| 37 |
% \date{2026, jan. 27} |
| 38 |
% |
| 39 |
% \maketitle |
| 40 |
% |
| 41 |
% \begin{abstract}\centerlastline\noindent |
| 42 |
% Sometimes one wants to get font size or letter spacing so that |
| 43 |
% text fits a desired length, or some unit length so that a picture |
| 44 |
% fits some desired size, for instance. This package implements |
| 45 |
% macros to compute such parameters to fit a condition. |
| 46 |
% \end{abstract} |
| 47 |
% |
| 48 |
% \section{The purpose of this package} |
| 49 |
|
| 50 |
% Suppose that one wants to create a nice logo with two lines of same |
| 51 |
% size, first line “AWESOME” and second line “example of use for |
| 52 |
% \texttt{paramcalc}”. We look for font size of first line, so that |
| 53 |
% the lines have the same width. This is the desired result: |
| 54 |
% |
| 55 |
% \newlength{\lineIwd} |
| 56 |
% \newlength{\lineIIwd} |
| 57 |
% \newcommand{\mytext}{\Large example of use for paramcalc} |
| 58 |
% \settowidth{\lineIIwd}{\mytext} |
| 59 |
% |
| 60 |
% \paramcalc{12}{50}{\lineIIwd}[\lineIwd]{ |
| 61 |
% \settowidth{\lineIwd}{\fontsize{\x}{\x}\selectfont \textls*{AWESOME}} |
| 62 |
% } |
| 63 |
% |
| 64 |
% \begin{center} |
| 65 |
% {\fontsize{\x}{\x}\selectfont \textls*{AWESOME}\par} |
| 66 |
% \mytext \vspace*{\bigskipamount} |
| 67 |
% \end{center} |
| 68 |
% |
| 69 |
% Instead of setting font size by hand\footnote{As width is |
| 70 |
% proportional to font size, one could also compute font size by |
| 71 |
% proportion, measuring texts of both lines.} and visually trying to get the |
| 72 |
% desired condition, we want to compute such font size for line 1. |
| 73 |
% |
| 74 |
% This can be acomplished with \texttt{paramcalc} with the following |
| 75 |
% code (make sure you use scalable fonts\footnote{Computer Modern are |
| 76 |
% not scalable, but Latin Modern (package \texttt{lmodern}), EB |
| 77 |
% Garamond (package \texttt{ebgaramond}) and many others are.}): |
| 78 |
% |
| 79 |
% \enlargethispage{\baselineskip} |
| 80 |
% \begin{verbatim} |
| 81 |
%\newcommand{\mytext}{\Large example of use for paramcalc} |
| 82 |
%\newlength{\lineIIwd} |
| 83 |
%\settowidth{\lineIIwd}{\mytext} |
| 84 |
% |
| 85 |
%\newlength{\lineIwd} |
| 86 |
%\paramcalc{12}{50}{\lineIIwd}[\lineIwd]{ |
| 87 |
% % compute \lineIwd as function of fontsize \x |
| 88 |
% \settowidth{\lineIwd}{\fontsize{\x}{\x}\selectfont \textls*{AWESOME}} |
| 89 |
%} |
| 90 |
% |
| 91 |
%Showing the result. |
| 92 |
%\begin{center} |
| 93 |
% {\fontsize{\x}{\x}\selectfont \textls*{AWESOME}\par} |
| 94 |
% \mytext |
| 95 |
%\end{center} |
| 96 |
%\end{verbatim} |
| 97 |
% |
| 98 |
% It you use \texttt{\string\x} in the document, you get the computed |
| 99 |
% value \x. |
| 100 |
% |
| 101 |
% \section{Usage} |
| 102 |
% |
| 103 |
% To compute a desired parameter, the situation can be reduced to find |
| 104 |
% a zero of a real function. In the example above, we wanted that the |
| 105 |
% length $y(x)$ of first line as function of its font size $x$ should |
| 106 |
% match the length $y_0$ of second line, so that we need to find the |
| 107 |
% zero of the function $f(x) = y(x) - y_0$. |
| 108 |
% |
| 109 |
% The computation is done by \texttt{\string\paramcalc} macro.\medskip |
| 110 |
% |
| 111 |
% \DescribeMacro{\paramcalc}^^A |
| 112 |
% \newcommand{\varp}[1]{\textlangle\makebox{\textit{#1}}\textrangle}^^A |
| 113 |
% \newcommand{\varm}[1]{\texttt{\{}\varp{#1}\texttt{\}}}^^A |
| 114 |
% \newcommand{\varopt}[1]{\texttt{[}\varp{#1}\texttt{]}} |
| 115 |
% \noindent\texttt{\string \paramcalc}\varopt{x macro}^^A |
| 116 |
% \varm{x initial}\varm{x final}\varopt{$\epsilon$}^^A |
| 117 |
% \varm{y desired}\varopt{y macro}\varm{body}\medskip |
| 118 |
% |
| 119 |
% The optional \varp{x macro} defaults to \texttt{\string\x}, the |
| 120 |
% optional \varp{y macro} defaults to \texttt{\string\y} and both |
| 121 |
% should be control sequences. \varp{y macro} and \varp{y desited} |
| 122 |
% could be lengths, converted to numbers as pt. |
| 123 |
% |
| 124 |
% \varp{body} should be code that computes \varp{y macro} for an |
| 125 |
% arbitrary value of \varp{x macro}, supposed to expand to a decimal |
| 126 |
% value. The result is stored in \varp{x macro} after its |
| 127 |
% computation. If its value is in between \varp{x initial} and |
| 128 |
% \varp{x final}, convergence is ensured. Often\footnote{I could not |
| 129 |
% figure out any real life example where the function was not affine.} |
| 130 |
% the functions involved are affine (for instance, a width is |
| 131 |
% proportional a font size), and in this case convergence is achieved |
| 132 |
% after just one step. |
| 133 |
% |
| 134 |
% The convergence is achieved when |
| 135 |
% $\vert \text{\varp{y macro}} - \text{\varp{y desired}}\vert < |
| 136 |
% \epsilon$. |
| 137 |
% The value of $\epsilon$, where $\epsilon$ is given in optional arg |
| 138 |
% \varp{epsilon} and defaults to 0.05. If it doesn't happen after a |
| 139 |
% maximum number of steps, an error is signaled. Then, probably de |
| 140 |
% desired $x$ is outside initial interval. |
| 141 |
% |
| 142 |
% There are situations where only integer values of $x$ are suitable, |
| 143 |
% like in the optional parameter \textit{n} for |
| 144 |
% \texttt{\string\textls}\varopt{n}\varm{text}. In such cases, use |
| 145 |
% \texttt{\string\intparamcalc}, with same arguments. The only |
| 146 |
% difference is that \varp{x macro} only assume integer values, but |
| 147 |
% \varp{y macro} still are supposed to assume real values.\medskip |
| 148 |
% |
| 149 |
% \DescribeMacro{\intparamcalc}^^A |
| 150 |
% \noindent\texttt{\string \intparamcalc}\varopt{x macro}^^A |
| 151 |
% \varm{x initial}\varm{x final}\varopt{$\epsilon$}^^A |
| 152 |
% \varm{y desired}\varopt{y macro}\varm{body}\medskip |
| 153 |
% |
| 154 |
% \textit{Attention}: The numerical method implemented in this package |
| 155 |
% is not suitable when $y$ assumes (too) discrete values or isn't a |
| 156 |
% continuous function of $x$ (it its where real-valued) like when $y$ |
| 157 |
% counts lines, for instance. |
| 158 |
% |
| 159 |
% \subsection{Package options} |
| 160 |
% |
| 161 |
% \begin{itemize} |
| 162 |
% \item \texttt{maxsteps} = \varp{n} : sets maximum number of steps |
| 163 |
% before giving up. Defaults to 20. |
| 164 |
% \item \texttt{epsilon} = \varp{epsilon} : sets the convegence |
| 165 |
% criterium $\epsilon$ to \varp{epsilon}. Defaults to 0.05. |
| 166 |
% \end{itemize} |
| 167 |
% |
| 168 |
% \section{Examples} |
| 169 |
% |
| 170 |
% Creating a logo for the book “História Geral do Brasil”, in three |
| 171 |
% lines. |
| 172 |
% \begin{itemize} |
| 173 |
% \item The first two lines in font size 20pt of size; |
| 174 |
% \item For the first line we compute letter spacing to fit size of 2nd line; |
| 175 |
% \item For the third line, we compute font size, keeping same size. |
| 176 |
% \end{itemize} |
| 177 |
% |
| 178 |
%\begin{center} |
| 179 |
% \fontsize{20}{22}\bfseries |
| 180 |
% |
| 181 |
% \newlength{\lIIwd} \newlength{\auxwd} |
| 182 |
% \settowidth{\lIIwd}{\textls*{GERAL \,DO}} |
| 183 |
% |
| 184 |
% \intparamcalc[\lIls]{0}{200}{\lIIwd}[\auxwd]{ |
| 185 |
% \settowidth{\auxwd}{\textls*[\lIls]{HISTÓRIA}}} |
| 186 |
% |
| 187 |
% \paramcalc[\lIIIfs]{20}{50}{\lIIwd}[\auxwd]{ |
| 188 |
% \settowidth{\auxwd}{\fontsize{\lIIIfs}{\lIIIfs}\selectfont \textls*{BRASIL}}} |
| 189 |
% |
| 190 |
% \textls*[\lIls]{HISTÓRIA} \\ \textls*{GERAL \,DO} \\ |
| 191 |
% \fontsize{\lIIIfs}{\lIIIfs}\selectfont \textls*{BRASIL} |
| 192 |
% \end{center} |
| 193 |
% |
| 194 |
% This is achieved with the code: |
| 195 |
% \begin{verbatim} |
| 196 |
%\begin{center} |
| 197 |
% \fontsize{20}{22}\bfseries |
| 198 |
% |
| 199 |
% % base length \lIIwd: line 2 with \textls* and \bfseries |
| 200 |
% \newlength{\lIIwd} \newlength{\auxwd} |
| 201 |
% \settowidth{\lIIwd}{\textls*{GERAL \,DO}} |
| 202 |
% |
| 203 |
% % compute letter spacing \lIls for 1st line: body compute \auxwd |
| 204 |
% \intparamcalc[\lIls]{0}{200}{\lIIwd}[\auxwd]{ |
| 205 |
% \settowidth{\auxwd}{\textls*[\lIls]{HISTÓRIA}}} |
| 206 |
% |
| 207 |
% % compute font size \lIIIfs for 3rd line: body compute \auxwd |
| 208 |
% \paramcalc[\lIIIfs]{20}{50}{\lIIwd}[\auxwd]{ |
| 209 |
% \settowidth{\auxwd}{\fontsize{\lIIIfs}{\lIIIfs}\selectfont \textls*{BRASIL}}} |
| 210 |
% |
| 211 |
% % typesetting the logo |
| 212 |
% \textls*[\lIls]{HISTÓRIA} \\ \textls*{GERAL \,DO} \\ |
| 213 |
% \fontsize{\lIIIfs}{\lIIIfs}\selectfont \textls*{BRASIL} |
| 214 |
%\end{center} |
| 215 |
% \end{verbatim} |
| 216 |
% |
| 217 |
% \section{Numerical method implemented} |
| 218 |
% |
| 219 |
% To ensure convergense (with a root in $[a,b]$), we implement Illinois |
| 220 |
% algorithm\footnote{Dowell, M., Jarratt, P. A modified regula falsi |
| 221 |
% method for computing the root of an equation. \textit{BIT Numerical |
| 222 |
% mathematics} \textbf{11}, 168–174 |
| 223 |
% (1971). \url{https://doi.org/10.1007/BF01934364}}, an improvement |
| 224 |
% over False Position method (\textit{regula falsi}): |
| 225 |
% |
| 226 |
% Assuming that $f:I\to \mathbb{R}$ is a continuous function, |
| 227 |
% end-points $a, b \in I$, $f(a)\cdot f(b) < 0$, so $f$ has a root in $[a,b]$. |
| 228 |
% |
| 229 |
% Let $c$ be the root of the secant line that connects $(a,f(a))$ and |
| 230 |
% $(b,f(b))$. |
| 231 |
% |
| 232 |
% \begin{center} |
| 233 |
% \begin{tikzpicture} |
| 234 |
% \draw[->] (-0.5,0) -- (2.5,0); \draw[->] (0,-1.5) -- (0,1.5); |
| 235 |
% \draw[thick] (0.5,-1) coordinate (a) |
| 236 |
% (a|-0,0) node[above] {$a$} |
| 237 |
% (1.8,1.5) coordinate (b) |
| 238 |
% (b|-0,0) node[below right] {$b$} |
| 239 |
% (a) -- (b) |
| 240 |
% (1.02,0) coordinate (x) node[below right] {$c$}; |
| 241 |
% \draw[dashed] (a|-0,0) -- (a) -- (a-|b) -- (b); |
| 242 |
% \fill (a) circle[radius=1pt] (b) |
| 243 |
% circle[radius=1pt] (x) circle[radius=1pt]; |
| 244 |
% \end{tikzpicture} |
| 245 |
% \end{center} |
| 246 |
% so, by similarity of triangles, |
| 247 |
% \begin{equation}\label{eq:c} |
| 248 |
% \Delta = \frac{b-a}{f(b)-f(a)} = \frac{b-c}{f(b)} \implies |
| 249 |
% c = b - f(b)\Delta. |
| 250 |
% \end{equation} |
| 251 |
% |
| 252 |
% The Illinois algorithm for False Position puts the following iteraction: |
| 253 |
% \begin{align} |
| 254 |
% (a,f(a)) & = |
| 255 |
% \begin{cases} |
| 256 |
% \bigl(c,f(c)\bigr), & \text{if } f(b)f(c)< 0,\\ |
| 257 |
% \bigl(a,\frac{f(a)}{2}\bigr), & \text{otherwise} |
| 258 |
% \end{cases}\label{eq:iter1} \\ |
| 259 |
% (b,f(b)) & = (c,f(c)) \label{eq:iter2} |
| 260 |
% \end{align} |
| 261 |
% |
| 262 |
% \section{Code} |
| 263 |
% |
| 264 |
% Identidication of the package. |
| 265 |
% \begin{macrocode} |
| 266 |
\NeedsTeXFormat{LaTeX2e} |
| 267 |
\ProvidesPackage{paramcalc}[2026/01/27 v1.0 compute params to fit conditions] |
| 268 |
% \end{macrocode} |
| 269 |
% |
| 270 |
% We use \LaTeX3 codebase. |
| 271 |
% \begin{macrocode} |
| 272 |
\ExplSyntaxOn |
| 273 |
% \end{macrocode} |
| 274 |
% |
| 275 |
% There are two versions of the solving macros: real roots are |
| 276 |
% admitted or only integer roots (like parameters for |
| 277 |
% \texttt{\string\textls[\string\x]\{\textit{text}\}}). We do a |
| 278 |
% single implementation and the cases are regulated by a boolen. |
| 279 |
% \begin{macrocode} |
| 280 |
\bool_new:N \g_PARC_int_bool |
| 281 |
% \end{macrocode} |
| 282 |
% |
| 283 |
% The iteraction will end when $|f(c)|<\epsilon$. |
| 284 |
% |
| 285 |
% Floating point variables for $a$, $b$, $f(a)$, $f(b)$, $c$, $f(c)$ |
| 286 |
% and $\Delta$: |
| 287 |
% \begin{macrocode} |
| 288 |
\fp_new:N \l_PARC_a_fp |
| 289 |
\fp_new:N \l_PARC_b_fp |
| 290 |
\fp_new:N \l_PARC_fa_fp |
| 291 |
\fp_new:N \l_PARC_fb_fp |
| 292 |
\fp_new:N \l_PARC_x_fp |
| 293 |
\fp_new:N \l_PARC_fx_fp |
| 294 |
\fp_new:N \l_PARC_delta_fp |
| 295 |
% \end{macrocode} |
| 296 |
% Also, a counter for steps taken, in case of divergence. |
| 297 |
% \begin{macrocode} |
| 298 |
\int_new:N \l_PARC_steps_int |
| 299 |
% \end{macrocode} |
| 300 |
% |
| 301 |
% Defining package option for maximum number of steps. |
| 302 |
% \begin{macrocode} |
| 303 |
\int_new:N \l_PARC_maxsteps_int |
| 304 |
\int_set:Nn \l_PARC_maxsteps_int { 20 } |
| 305 |
\keys_define:nn {paramcalc} { maxsteps .int_set:N = \l_PARC_maxsteps_int } |
| 306 |
% \end{macrocode} |
| 307 |
% |
| 308 |
% Defining package option for default epsilon (stop criterium). |
| 309 |
% \begin{macrocode} |
| 310 |
\fp_new:N \l_PARC_epsilon_fp |
| 311 |
\fp_set:Nn \l_PARC_epsilon_fp { 0.05 } |
| 312 |
\keys_define:nn {paramcalc} { epsilon .fp_set:N = \l_PARC_epsilon_fp } |
| 313 |
% \end{macrocode} |
| 314 |
% |
| 315 |
% Passing options to package. |
| 316 |
% \begin{macrocode} |
| 317 |
\DeclareOption*{ \keys_set:ne { paramcalc } { \CurrentOption } } |
| 318 |
\ProcessOptions |
| 319 |
% \end{macrocode} |
| 320 |
% |
| 321 |
% \DescribeMacro{\PARC\_paramcalc:NnnnnNn} |
| 322 |
% The arguments of $\mathtt{\string\PARC\_paramcalc:NnnnnNn}$ |
| 323 |
% are\\ |
| 324 |
% \texttt{\#1} = macro that holds value for $x$ (default to |
| 325 |
% \texttt{\string\x} later on)\\ |
| 326 |
% \texttt{\#2} = $a$\\ |
| 327 |
% \texttt{\#3} = $b$\\ |
| 328 |
% \texttt{\#4} = $\epsilon$\\ |
| 329 |
% \texttt{\#5} = $y_0$\\ |
| 330 |
% \texttt{\#6} = macro that stores computed $y(x)$ (default to |
| 331 |
% \texttt{\string\y} later on)\\ |
| 332 |
% \texttt{\#7} = body of code that computes $y(x)$ and stores value in |
| 333 |
% \texttt{\#6}. |
| 334 |
% \begin{macrocode} |
| 335 |
\cs_new:Nn \PARC_paramcalc:NnnnnNn |
| 336 |
{ |
| 337 |
% \end{macrocode} |
| 338 |
% Store $a$ from \texttt{\#2} in \var{l_PARC_a_fp} and set |
| 339 |
% \texttt{\#1} as the decimal representation of $a$, taking care if |
| 340 |
% integer values must be used. |
| 341 |
% \begin{macrocode} |
| 342 |
\fp_set:Nn \l_PARC_a_fp { #2 } |
| 343 |
\bool_if:NTF \g_PARC_int_bool |
| 344 |
{ \tl_set:Ne #1 { \fp_to_int:n { #2 } } } |
| 345 |
{ \tl_set:Ne #1 { \fp_to_decimal:n { #2 } } } |
| 346 |
% \end{macrocode} |
| 347 |
% Call body to get $y(x)$ in \texttt{\#6} |
| 348 |
% \begin{macrocode} |
| 349 |
#7 |
| 350 |
% \end{macrocode} |
| 351 |
% Save $f(x) = y(x) - y_0$ in \var{l_PARC_fa_fp} |
| 352 |
% \begin{macrocode} |
| 353 |
\fp_set:Nn \l_PARC_fa_fp { #6 - (#5) } |
| 354 |
% \end{macrocode} |
| 355 |
% |
| 356 |
% Do as before, now for $b$ stored in \texttt{\#3}: |
| 357 |
% \begin{macrocode} |
| 358 |
\fp_set:Nn \l_PARC_b_fp { #3 } |
| 359 |
\bool_if:NTF \g_PARC_int_bool |
| 360 |
{ \tl_set:Ne #1 { \fp_to_int:n { #3 } } } |
| 361 |
{ \tl_set:Ne #1 { \fp_to_decimal:n { #3 } } } |
| 362 |
#7 |
| 363 |
\fp_set:Nn \l_PARC_fb_fp { #6 - (#5) } |
| 364 |
% \end{macrocode} |
| 365 |
% |
| 366 |
% In many situations, the function $f(x)$ is affine and the secant |
| 367 |
% method converges after the first iteration, so it really doesn't |
| 368 |
% matter if $f(a)f(b)<0$, so we trigger just a warning if signs of |
| 369 |
% $f(a)$ and $f(b)$ are the same. |
| 370 |
% \begin{macrocode} |
| 371 |
\fp_compare:nNnT { (\l_PARC_fa_fp) * (\l_PARC_fb_fp) } > { 0 } |
| 372 |
{ |
| 373 |
\msg_warning:nn {paramcalc} |
| 374 |
{ |
| 375 |
The ~values ~of ~\string#6 ~are |
| 376 |
~{\fp_to_decimal:N \l_PARC_fa_fp} ~and |
| 377 |
~{\fp_to_decimal:N \l_PARC_fa_fp}, ~both |
| 378 |
\fp_compare:nNnTF { \l_PARC_fa_fp } > 0 { ~above } { ~below } |
| 379 |
~{\fp_to_decimal:n {#5}} ~so ~convergence ~is ~not ~ensured.\\ |
| 380 |
Try ~to ~set ~initial ~values ~that ~surround ~the ~root. |
| 381 |
} |
| 382 |
} |
| 383 |
% \end{macrocode} |
| 384 |
% \bigskip |
| 385 |
% |
| 386 |
% Now the loop. Last iteraction is supposed to be stored in |
| 387 |
% \var{l_PARC_fb_fp}. Stop condition is $|f(b)|< \epsilon$, with |
| 388 |
% $\epsilon$ stored in \texttt{\#4}. |
| 389 |
% \begin{macrocode} |
| 390 |
\fp_while_do:nNnn { abs(\l_PARC_fb_fp) } > { #4 } |
| 391 |
{ |
| 392 |
% \end{macrocode} |
| 393 |
% Computing $\Delta$ and $c$ as in \eqref{eq:c}. |
| 394 |
% \begin{macrocode} |
| 395 |
\fp_set:Nn \l_PARC_delta_fp |
| 396 |
{ (\l_PARC_b_fp - \l_PARC_a_fp) / (\l_PARC_fb_fp - \l_PARC_fa_fp) } |
| 397 |
\fp_set:Nn \l_PARC_c_fp |
| 398 |
{ \l_PARC_b_fp - (\l_PARC_fb_fp) * (\l_PARC_delta_fp) } |
| 399 |
% \end{macrocode} |
| 400 |
% Compute $f(c)$ and store it in \var{l_PARC_fc_fp}. |
| 401 |
% \begin{macrocode} |
| 402 |
\bool_if:NTF \g_PARC_int_bool |
| 403 |
{ \tl_set:Ne #1 { \fp_to_int:N \l_PARC_c_fp } } |
| 404 |
{ \tl_set:Ne #1 { \fp_to_decimal:N \l_PARC_c_fp } } |
| 405 |
#7 |
| 406 |
\fp_set:Nn \l_PARC_fc_fp { #6 - (#5) } |
| 407 |
% \end{macrocode} |
| 408 |
% Now implement Illinois iteraction from \eqref{eq:iter1} |
| 409 |
% \begin{macrocode} |
| 410 |
\fp_compare:nNnTF { (\l_PARC_fc_fp) * (\l_PARC_fc_fp) } < { 0 } |
| 411 |
{ |
| 412 |
\fp_set:Nn \l_PARC_a_fp { \l_PARC_b_fp } |
| 413 |
\fp_set:Nn \l_PARC_fa_fp { \l_PARC_fb_fp } |
| 414 |
} |
| 415 |
{ |
| 416 |
\fp_set:Nn \l_PARC_fa_fp { 0.5 * (\l_PARC_fa_fp) } |
| 417 |
} |
| 418 |
% \end{macrocode} |
| 419 |
% and \eqref{eq:iter2}. |
| 420 |
% \begin{macrocode} |
| 421 |
\fp_set:Nn \l_PARC_b_fp { \l_PARC_c_fp } |
| 422 |
\fp_set:Nn \l_PARC_fb_fp { \l_PARC_fc_fp } |
| 423 |
} |
| 424 |
% \end{macrocode} |
| 425 |
% |
| 426 |
% Handling possible divergence if $f(a)f(b)>0$. If it doesn't converge |
| 427 |
% after \texttt{\string\l\_PARC\_maxsteps\_int} steps, an error is |
| 428 |
% signaled. |
| 429 |
% \begin{macrocode} |
| 430 |
\int_incr:N \l_PARC_steps_int |
| 431 |
\int_compare:nNnT { \l_PARC_steps_int } > { \l_PARC_maxsteps_int } |
| 432 |
{ |
| 433 |
\msg_error:nn {paramcalc} |
| 434 |
{ |
| 435 |
Failed ~to ~converge ~after ~\int_use:N \l_PARC_steps_int ~steps.\\ |
| 436 |
Try ~to ~narrow ~initial ~values ~around ~the ~zero. |
| 437 |
} |
| 438 |
} |
| 439 |
} |
| 440 |
% \end{macrocode} |
| 441 |
% Recall that the user access the computed value from \texttt{\#1}. |
| 442 |
% |
| 443 |
% Now we define user functions that set \var{g_PARC_int_bool} and |
| 444 |
% call previous function. |
| 445 |
% |
| 446 |
% Define \texttt{\string\paramcalc}. |
| 447 |
% \begin{macrocode} |
| 448 |
\NewDocumentCommand{\paramcalc}{ O{\x} m m O{0.05} m O{\y} m} |
| 449 |
{ |
| 450 |
\bool_set:Nn \g_PARC_int_bool { \c_false_bool } |
| 451 |
\PARC_paramcalc:NnnnnNn #1 { #2 } { #3 } { #4 } { #5 } #6 { #7 } |
| 452 |
} |
| 453 |
% |
| 454 |
% Define \texttt{\string\intparamcalc}. |
| 455 |
% \begin{macrocode} |
| 456 |
\NewDocumentCommand{\intparamcalc}{ O{\x} m m O{0.05} m O{\y} m} |
| 457 |
{ |
| 458 |
\bool_set:Nn \g_PARC_int_bool { \c_true_bool } |
| 459 |
\PARC_paramcalc:NnnnnNn #1 { #2 } { #3 } { #4 } { #5 } #6 { #7 } |
| 460 |
} |
| 461 |
% \end{macrocode} |
| 462 |
% |
| 463 |
% Finish \LaTeX3 syntax. |
| 464 |
% \begin{macrocode} |
| 465 |
\ExplSyntaxOff |
| 466 |
% \end{macrocode} |
| 467 |
% |
| 468 |
% \Finale |
| 469 |
% |
| 470 |
\endinput |
| 471 |
% |
| 472 |
%% \CharacterTable |
| 473 |
%% {Upper-case \A\B\C\D\E\F\G\H\I\J\K\L\M\N\O\P\Q\R\S\T\U\V\W\X\Y\Z |
| 474 |
%% Lower-case \a\b\c\d\e\f\g\h\i\j\k\l\m\n\o\p\q\r\s\t\u\v\w\x\y\z |
| 475 |
%% Digits \0\1\2\3\4\5\6\7\8\9 |
| 476 |
%% Exclamation \! Double quote \" Hash (number) \# |
| 477 |
%% Dollar \$ Percent \% Ampersand \& |
| 478 |
%% Acute accent \' Left paren \( Right paren \) |
| 479 |
%% Asterisk \* Plus \+ Comma \, |
| 480 |
%% Minus \- Point \. Solidus \/ |
| 481 |
%% Colon \: Semicolon \; Less than \< |
| 482 |
%% Equals \= Greater than \> Question mark \? |
| 483 |
%% Commercial at \@ Left bracket \[ Backslash \\ |
| 484 |
%% Right bracket \] Circumflex \^ Underscore \_ |
| 485 |
%% Grave accent \` Left brace \{ Vertical bar \| |
| 486 |
%% Right brace \} Tilde \~} |