10

If I use mathpazo then it seems that \left\Vert and \right\Vert generate the wrong size in one case:

 \documentclass[11pt]{article}
 \usepackage{mathpazo}
 \begin{document}
 {\Large $\left\Vert X_f\right\Vert$}
 {\large $\left\Vert X_f\right\Vert$}
 {\normalsize $\left\Vert X_f\right\Vert$}
 {\small $\left\Vert X_f\right\Vert$}
 {\footnotesize $\left\Vert X_f\right\Vert$}
 {\scriptsize $\left\Vert X_f\right\Vert$}
 {\tiny $\left\Vert X_f\right\Vert$}
 \end{document}

Notice how the fourth \Vert is the wrong size:

enter image description here

Is this a bug or a feature? And how to I get rid of it? Saying "do not use \left and \right" is not a valid answer until you also suggest how I might get the following macro working without them:

\newcommand{\trunc}[2]{\mathopen{}\left\Vert #2\right\Vert_{#1}\mathclose{}}

It has to be one macro, the book is already written using it. As far as I am concerned this is not my fault, but that of mathpazo.

Andrej Bauer
  • 252
  • How to get rid of it? Do not use \left and \right, they produce an extra horizontal space and they quite often produce too large delimiters. – yo' Jun 10 '13 at 22:13
  • Ok but what should I use then? This is for a macro which has to automagically pick the correct size of \Vert depending on the size of the argument. Let me amend the question. – Andrej Bauer Jun 10 '13 at 22:18
  • By the way, this is not something special about \Vert -- it applies to all other delimiters too, at least all the ones I've tried. – Mike Shulman Jun 10 '13 at 23:11
  • in my opinion, only the first three and the next-to-last look decent. using \smash[b] on only the subscript (requires amsmath) would be an improvement, but it may not be practical. i'll take a closer look, with other delimiters as well as \Vert and parentheses (as in your original formulation). – barbara beeton Jun 10 '13 at 23:18
  • You'll find a solution in http://tex.stackexchange.com/questions/102157/sloppy-left-and-right. The commands you want are \delimitershortfall=10pt and \delimiterfactor=750. This is not just a problem with mathpazo, try compiling with \usepackage{mathpazo} removed. Then the scriptsize and tiny delimeters are poorly sized. Compiling the document at 10,11 and 12pt produces surprising results, with mathpazo and also without mathpazo. – corporal Jun 11 '13 at 01:16
  • It seems to be a specific problem with the f subscript - replacing it with g does not display such drastic behaviour. – Andrew Swann Jun 11 '13 at 07:57
  • 2
    I don't suppose anyone would be interested in explaining what is causing this problem, and why \delimiterfactor and \delimitershortfall fix it? Personally, I would really like to understand, not just slap on a quick fix that acts like magic to me. – Mike Shulman Jun 11 '13 at 14:43

1 Answers1

3

You can try the nath or the MnSymbol packages. Beware that they have to be loaded after all packages that define math stuff (and might be incompatible with some of them).

Your MWE (I replaced \left\Vert with \lVert and \right\Vert with \rVert)

\documentclass[11pt]{article}
\usepackage{mathpazo}
\usepackage{nath}
\begin{document}
{\Large $\lVert X_f\rVert$}
{\large $\lVert X_f\rVert$}
{\normalsize $\lVert X_f\rVert$}
{\small $\lVert X_f\rVert$}
{\footnotesize $\lVert X_f\rVert$}
{\scriptsize $\lVert X_f\rVert$}
{\tiny $\lVert X_f\rVert$}
\end{document}

with nath package, for example, will produce

enter image description here

A similar result is obtained with MnSymbol.

karlkoeller
  • 124,410