Take the 2-minute tour ×
TeX - LaTeX Stack Exchange is a question and answer site for users of TeX, LaTeX, ConTeXt, and related typesetting systems. It's 100% free, no registration required.

I would like to construct what Unicode calls "white curly brackets" that are in addition scalable. They look like ⦃ (≈{|) and ⦄ (≈|}) and have Unicode codepoints U+2983 and U+2984.

unicode-math has them as \lBrace and \rBrace, but (1) I don't presently use XeTeX or LuaTeX and (2) their design (shown in unimath-symbols.pdf on pages 3-4) is not to my liking: the brace part is too fat for my taste, and I would like the vertical bar to not leave any protrusions. I am trying to have them visually match the parentheses \llparenthesis, \rrparenthesis, \llbracket, and \rrbracket from stmaryrd.

Right now I am using a makeshift definition, which is sufficient for now, but the symbols don't scale (in general or with \left/\right). Here is sample code:

\documentclass{article}
\usepackage{amsmath}
\usepackage[only,llbracket,rrbracket,llparenthesis,rrparenthesis]{stmaryrd} % for comparison with 4 similar parenthesis symbols
\usepackage{accsupp} % for ensuring the right Unicode codepoint upon pasting

\newcommand*{\llbrace}{%
  \BeginAccSupp{method=hex,unicode,ActualText=2983}%
    \textnormal{\usefont{OMS}{lmr}{m}{n}\char102}%
    \mathchoice{\mkern-4.05mu}{\mkern-4.05mu}{\mkern-4.3mu}{\mkern-4.8mu}%
    \textnormal{\usefont{OMS}{lmr}{m}{n}\char106}%
  \EndAccSupp{}%
}
\newcommand*{\rrbrace}{%
  \BeginAccSupp{method=hex,unicode,ActualText=2984}%
    \textnormal{\usefont{OMS}{lmr}{m}{n}\char106}%
    \mathchoice{\mkern-4.05mu}{\mkern-4.05mu}{\mkern-4.3mu}{\mkern-4.8mu}%
    \textnormal{\usefont{OMS}{lmr}{m}{n}\char103}%
  \EndAccSupp{}%
}


\begin{document}

\noindent
ok:
\(\displaystyle\llbrace abc \rrbrace\)
\(\textstyle\llbrace abc \rrbrace\)
\(\scriptstyle\llbrace abc \rrbrace\)
\(\scriptscriptstyle\llbrace abc \rrbrace\)

\noindent
ok:
\(\displaystyle \llbrace \frac{a}{b} \rrbrace\)
\(\textstyle\llbrace \frac{a}{b} \rrbrace\)
\(\scriptstyle\llbrace \frac{a}{b} \rrbrace\)
\(\scriptscriptstyle\llbrace \frac{a}{b} \rrbrace\)

\noindent
incorrectly scaled (\verb|\scriptsize|): {\scriptsize\(\llbrace abc \rrbrace\)} (large magnification required)

\noindent
incorrectly scaled (\verb|\tiny|): {\tiny\(\llbrace abc \rrbrace\)}

\noindent
\verb|\left| and \verb|\right| not possible before \verb|\llbrace| and \verb|\rrbrace|: \\
\indent \(\llbrace \left\llbracket\left\{\left[(a) \cdot bc^{2^3}\right]^4\right\}^5\right\rrbracket \rrbrace\)
% Note: This is also not possible with \llparenthesis and \rrparenthesis from stmaryrd.

\noindent
for comparison: \(\llparenthesis x \rrparenthesis, \llbracket x \rrbracket, \llbrace x \rrbrace; (x), [x], \{x\}\)

\end{document}

Lines 1 and 2 look just right with the present setup. Lines 3 and 4 demonstrate that my definitions don't scale. Lines 5a/b gives an example of where I would like to be able to use \left and \right. Line 6 juxtaposes different types of parentheses for visual comparison.

Is there an easy way to scale the glyphs I have constructed myself? Or is designing new glyphs the only way? While the vertical part of my glyphs doesn't leave any protrusions, the "brace part being too fat" part of my problem isn't being addressed by my code. With this in mind: I recognize that it might not be that easy to make the brace part of the glyph thinner (though if you can, more power to you), but I think that any solution that addresses the scaling problem for any manually constructed glyph cluster will be a valuable contribution to the community.


For a slightly different setup (original example code) with

\RequirePackage{fix-cm}
\documentclass[12pt]{memoir}
\usepackage{fixltx2e}[2006/09/13]
\usepackage{amsmath}
\usepackage{txfonts}
\usepackage[only,llbracket,rrbracket,llparenthesis,rrparenthesis]{stmaryrd}
\usepackage{accsupp}

in the preamble, one needs \mkern-5.7mu in the fourth argument to \mathchoice for "correct" scaling. To see what goes wrong in that case, it's best to include a line such as discontinuous small symbols in a footnote\footnote{\(\scriptscriptstyle\llbrace \frac{a}{b} \rrbrace\)}. This might help a potential answerer test things out.

share|improve this question
1  
If I look at all the packages you use: this is not what I'd call an MWE :-) Or are there particular reasons why all these packages are relevant to your problem? –  Hendrik Vogt Mar 5 '13 at 7:34
2  
@HendrikVogt After spending an hour playing with the \mkern-..mu-values in my present document, you really think I should try recalibrating things just to be able to write \documentclass{article}? memoir is a tasteful class, and using fix-cm and fixltx2e is good practice anyways (this can be relevant depending on the font tricks you're using). Without amsmath the glyph positioning already looks different (try it out). [12pt] can be done away with (again, at the cost of different letter positioning), but I don't think there is value in omitting this now just for saving 4 letters. –  Lover of Structure Mar 5 '13 at 7:56
    
@HendrikVogt But if you want to edit out accsupp (using it here would be useful and good practice) or other lines, I very much and strongly encourage you to make the necessary constructive edits ... As indicated, the sample code is showing one possible way to get the two symbols approximately right. A possible solution that manages to scale just those two symbols will or should be independent anyways; and if the only solution is to stick together glyphs in a totally different way, this will be so even more so. –  Lover of Structure Mar 5 '13 at 8:07
1  
Regarding accsupp: It's definitely not relevant to your problem, but I'm happy that I learned from you about that package and its usage :-) Still, the point about an MWE is not to show good practice. You can omit the fixes, the txfonts, and you can also replace memoir with article - no need to recalibrate anything, just try it! –  Hendrik Vogt Mar 5 '13 at 8:53
    
@HendrikVogt Okay, I will try :-) Gimme a few minutes ... –  Lover of Structure Mar 5 '13 at 8:55
show 2 more comments

2 Answers

up vote 8 down vote accepted

With scalerel, you can definitely take care of the \left \right problem. And I think I took care of the tine & scriptsize problem, too. What I did was to save a good copy of the normalsize bbraces in boxes, so that \lxbrace and \rxbrace would place the boxes containing good copies of the bbraces. I then used scalerel to scale these boxes to the desired sizes. So, essentially \lxbrace and \rxbrace replaced your \llbrace and \rrbrace.

The key here is that I am scaling the \normalsize version of the bbraces, whereas what I think happened to your original glyph is that tiny versions of a LaTeX brace are not merely scaled versions of the normalsize brace, but modified (by LaTeX) for legibility. But inyour case, you don't want modifications to your glyph's components as they are shrunk.

Lastly, scalerel is able to emulate the \left \right thingy to scale your brace up to the proper size, whatever that might be. I did use a width-limited scale, to avoid the scaled brace looking too clunky.

\documentclass{article}
\usepackage{amsmath}
\usepackage{scalerel}
\usepackage[only,llbracket,rrbracket,llparenthesis,rrparenthesis]{stmaryrd}
\parskip 1ex
\parindent 0ex
\usepackage{accsupp} % for ensuring the right Unicode codepoint upon pasting

\newcommand*{\llbrace}{%
  \BeginAccSupp{method=hex,unicode,ActualText=2983}%
    \textnormal{\usefont{OMS}{lmr}{m}{n}\char102}%
    \mathchoice{\mkern-4.05mu}{\mkern-4.05mu}{\mkern-4.3mu}{\mkern-4.8mu}%
    \textnormal{\usefont{OMS}{lmr}{m}{n}\char106}%
  \EndAccSupp{}%
}
\newcommand*{\rrbrace}{%
  \BeginAccSupp{method=hex,unicode,ActualText=2984}%
    \textnormal{\usefont{OMS}{lmr}{m}{n}\char106}%
    \mathchoice{\mkern-4.05mu}{\mkern-4.05mu}{\mkern-4.3mu}{\mkern-4.8mu}%
    \textnormal{\usefont{OMS}{lmr}{m}{n}\char103}%
  \EndAccSupp{}%
}

\begin{document}
\newsavebox{\lXbrace}
\savebox{\lXbrace}{$\llbrace$}
\newsavebox{\rXbrace}
\savebox{\rXbrace}{$\rrbrace$}
\def\lxbrace{\scalerel*{\usebox{\lXbrace}}{\llbrace}}
\def\rxbrace{\scalerel*{\usebox{\rXbrace}}{\rrbrace}}

\noindent
\verb|\tiny|: {\tiny\(\llbrace abc \rrbrace\)}
{\tiny\(\lxbrace abc \rxbrace\)} 
 $<-$scale of original brace

  \verb,ORIGINAL:,%
\indent \(\llbrace \left\llbracket\left\{\left[(a) \cdot
bc^{2^3}\right]^4\right\}^5\right\rrbracket \rrbrace\)
\\\verb,FIXED:   ,%
\def\core{ \left\llbracket\left\{\left[(a) \cdot
bc^{2^3}\right]^4\right\}^5\right\rrbracket}%
\indent \(\scalerel[1.5ex]{\lxbrace}{\core}\scalerel*[1.5ex]{\rxbrace}{\core}\)
bbraces scaled, but limited to 1.5ex width

\end{document}

I edited the original image (via zoom) to show how scalerel can take the bbraces and vertically scale them to fit the mathematical object they enclose. The top equation is without scalerel, the bottom with it.

enter image description here

I have added also a zoom on the \tiny size objects, so that one can see the effect of scaling the normalsize glyph rather than letting LaTeX scale both the strut and the brace and then sticking them together. The left version is letting LaTeX scale the glyphs that compose the character, while the right image is letting scalerel package scale the \normalsize character to \tiny proportions.

enter image description here


Based on request to be able to provide horizontal compression, to help make glyphs better match comparable counterparts in the literature, I upgraded the scalerel package today (to v1.3), adding the following composite commands

\newcommand\scaleleftright[4][99in]{%
  \if.#2\def\leftarg{\bl@nk}\else\def\leftarg{#2}\fi%
  \if.#4\def\rightarg{\bl@nk}\else\def\rightarg{#4}\fi%
  \scalerel[#1]{\leftarg}{#3}\scalerel*[#1]{\rightarg}{#3}%
}
\newcommand\stretchleftright[4][10000]{%
  \if.#2\def\leftarg{\bl@nk}\else\def\leftarg{#2}\fi%
  \if.#4\def\rightarg{\bl@nk}\else\def\rightarg{#4}\fi%
  \stretchrel[#1]{\leftarg}{#3}\stretchrel*[#1]{\rightarg}{#3}%
}

\newcommand\hstretch[2]{\stretchrel*{\scalebox{#1}{$#2$}}{#2}}

\newcommand\vstretch[2]{\stretchrel*{#2}{\scalebox{#1}{$#2$}}}

The ...leftright commands provide automated scaling for symbols in the manner of mathmode's \left and \right commands. But the \hstretch and \vstretch commands allow for the glyph stretching that was desired by the user. Using his prior example, I further refined the definition of his white-curly-braces symbol as

\def\lxbrace{%
   \hstretch{0.6}{\scalerel*{\usebox{\lXbrace}}{\llbrace}}}
\def\rxbrace{%
   \hstretch{0.6}{\scalerel*{\usebox{\rXbrace}}{\rrbrace}}}

which provided a 40% horizontal compression of his composite-glyph character. Thus, to compare (now in normal, not tiny size), my original correction to this latest correction:

enter image description here

share|improve this answer
    
What an impressive solution! Your scalerel package seems perfect for my problem. –  Lover of Structure Mar 16 '13 at 6:48
1  
@LoverofStructure I updated the scalerel package today (v1.2) to add \scaleleftright and \stretchleftright composite commands to mimic the \left ... \right functionality of math mode, since not all symbols will scale using that syntax of math mode. This isn't new functionality, rather just convenience of syntax. –  Steven B. Segletes Mar 18 '13 at 12:32
    
In your documentation, you ask for ideas about horizontal scaling. This isn't really an implementation idea, but as an idea for an application: one could make the curly portion of the "white curly brackets" shown/constructed here a bit thinner so that they match up better with the other two parenthesis pairs from stmaryrd that I'm using. At least for ordinary font sizes this difference is noticeable. For larger font sizes one could then horizontally stretch them again (with a limit, like what you already have). –  Lover of Structure Mar 20 '13 at 1:09
add comment

A poor man's approach. It probably doesn't fit the bill:

\documentclass{article}

\begin{document}

\Huge

$\left\lbrace\mkern-2mu\middle| ABC^{\left\lbrace\mkern-2mu\middle| XYZ\middle|\mkern-2mu\right\rbrace} \middle|\mkern-2mu\right\rbrace$

\end{document}

braces

share|improve this answer
1  
Thanks for your contribution! Actually I think your answer is an example of what doesn't work :-) (really no offense), so if you could edit your answer to (1) say this and (2) state what your approach is, I think it deserves an upvote for its illustrative value. (I will then also delete this comment.) –  Lover of Structure Mar 12 '13 at 4:49
add comment

Your Answer

 
discard

By posting your answer, you agree to the privacy policy and terms of service.

Not the answer you're looking for? Browse other questions tagged or ask your own question.