# Breaking up long equation with lots of parentheses — can't get rid of errors

I have read all answers that use align, multline, multlined etc. but I cannot get rid of errors in my case. Also, the use of \left. and \right. does not eliminate errors. Any help appreciated. The use of automatic \left and \right generates horribly high left parenthesis and anemic right ones, therefore all the visual pairing is gone down the drain, so I posted a second equation using \Biggl and \Biggr, which looks a lot better IMHO. But the errors from breaking the line in two are still there.

\documentclass{article}

\usepackage{amsmath}
\usepackage{mathtools}
\usepackage{amssymb}
\usepackage{color}

%lower the subscript by ex points
\def\dc{\rule{0pt}{1.0ex}}
\def\dd{\rule{0pt}{1.5ex}}
\def\df{\rule{0pt}{1.8ex}}
\def\ds{\rule{0pt}{2.0ex}}
\def\dss{\rule{0pt}{2.5ex}}
\def\dds{\rule{0pt}{3.0ex}}

\begin{document}

This is a very long equation...
\\
%\begin{align*}
\begin{multline*}
\leqslant\;
\sum_{j \,=\, 1}^{m}
%\left\Bigg\{\,
\left\{\,
\sum_{i \,=\, 1}^{n}
%\Bigg[\,
\left[\,
\int\displaylimits_{\ds{Z_{\dc{i}}}}{
%\Bigg(\,
\left(\,
%
% first norm
\Big\Vert\,
K
\Big(\,
\mu
\big(\,
\omega(A_{\dd{j}})
\,\big)
\,\Big)
\;-\;
K_{i}
\Big(\,
\mu
\big(\,
\omega(A_{\dd{j}})
\,\big)
\,\Big)
\,\Big\Vert
% end first norm
\;+\;
\right.
\right.
\right.
\\
\left.
\left.
\left.
%
% second norm
\Big\Vert\,
K
\Big(\,
\mu
\big(\,
\omega(A_{\dd{j}})
\,\big)
\,\Big)
\;-\;
K_{i}
\Big(\,
\mu
\big(\,
\omega(A_{\dd{j}})
\,\big)
\,\Big)
\,\Big\Vert
% end second norm
\,\right)
\,\mathrm{d}\lambda}
%\,\Bigg]
\,\right]
%\,\Bigg\}
\,\right\}
\;\leqslant\;
%\end{align*}
\end{multline*}
\\%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%

\begin{align*}
\leqslant\;
\sum_{j \,=\, 1}^{m}
\Biggl\{\,
\sum_{i \,=\, 1}^{n}
\Biggl[\,
\int\displaylimits_{\ds{Z_{\dc{i}}}}{
\Biggl(\,
%
% first norm
\Big\Vert\,
K
\Big(\,
\mu
\big(\,
\omega(A_{\dd{j}})
\,\big)
\,\Big)
\;-\;
K_{i}
\Big(\,
\mu
\big(\,
\omega(A_{\dd{j}})
\,\big)
\,\Big)
\,\Big\Vert
% end first norm
\;+\;
\right.
\right.
\right.
\\
\left.
\left.
\left.
%
% second norm
\Big\Vert\,
K
\Big(\,
\mu
\big(\,
\omega(A_{\dd{j}})
\,\big)
\,\Big)
\;-\;
K_{i}
\Big(\,
\mu
\big(\,
\omega(A_{\dd{j}})
\,\big)
\,\Big)
\,\Big\Vert
% end second norm
\,\Biggr)
\,\mathrm{d}\lambda}
\,\Biggr]
\,\Biggr\}
\;\leqslant\;
\end{align*}

\end{document}

• (1) why all those manual spacing commands? (2) all those left right constructions are useless anyway because even split using the dot construction you are likely to get un even scaling. My advice would be to start by removing all those left right constructions and go from there. I cannot test your code as I'm on a tablet. BTW it will be helpful to others if you cite the error you get. – daleif Sep 20 '15 at 15:28
• BTW with all this manual adjusting you are doing, you are really going to piss off journal editors who has to make sure things are up to the journal standard. – daleif Sep 20 '15 at 15:30
• You are starting from the false premise that you have to say \int\displaylimits_{<domain>}{<integrand>}. You should use \limits and no braces around the integrand. Of course the result, due to all those \big and spaces is completely wrong, from a typographic point of view. – egreg Sep 20 '15 at 15:46
• @egreg I tried \limits first, it did not align the expression equally/simetrically under the integral sign, instead it put all of it to the lower right of the integral. \displaylimits_ seems to do a better job for some reason, especially when you have under the integral a larger expression, like [ a, b ] for example. Thank you for the tip with the integrand, that was the problem really! I took out those curly braces around the integrand and all errors went away. – Diegis Sep 20 '15 at 15:53
• Since you are in a display, \limits and \displaylimits behave in exactly the same way. But that's not the main problem, just a question of style. The syntax error disappeared, yes, but the typesetting ones didn't. Please, leave TeX most of the math typesetting and don't be too generous with delimiter sizes. – egreg Sep 20 '15 at 16:01

## 1 Answer

The main error is in using

\int\displaylimits_{<subscript>}{<integrand>}


where the braces around the integrand are useless (and are the cause of your troubles.

Also \limits is sufficient and, of course, all that manual sizing and spacing produces a bad result.

In order to have openings you have to use \bigl and for closing \bigr (with the obvious variants). Moreover you're choosing too big sizes.

\documentclass{article}

\usepackage{amsmath}
\usepackage{mathtools}
\usepackage{amssymb}

\begin{document}

This is a very long equation...
\begin{multline*}
\leqslant
\sum_{j=1}^{m}
\biggl\{\, % this \, is good, because of \sum following
\sum_{i=1}^{n}
\biggl[\, % this \, is good, because of \int following
\int\limits_{Z_i}
\Bigl(
% first norm
\bigl\Vert
K\bigl(\mu(\omega(A_{j}))\bigr)-K_{i}\bigl(\mu(\omega(A_{j}))\bigr)
\bigr\Vert
% end first norm
+{}\\
% second norm
\bigl\Vert
K\bigl(\mu(\omega(A_{j}))\bigr)-K_{i}\bigl(\mu(\omega(A_{j}))\bigr)
\bigr\Vert
% end second norm
\Bigr) \,\mathrm{d}\lambda
\biggr]
\biggr\}
\leqslant
\end{multline*}

\end{document}


There is no reason for the fences to fully cover what they delimit.

• is there a reason you use \bigl\Vert and \bigr\Vert rather than \big\lVert and \big\rVert? (the results are the same, as they appear on my screen.) – barbara beeton Sep 20 '15 at 16:48
• @barbarabeeton No, they are not the same: with \bigl\Vert I get an Open atom; with \big\lVert I get an Ord atom. It's similar to \left\|, which is the same (but simpler to type) as \left\lVert`. – egreg Sep 20 '15 at 18:06
• Why is there a need for +{}\\ instead of simply +\\? The +\\ compiles fine and it looks exactly the same as +{}\\ on my machine... – Diegis Sep 20 '15 at 23:11
• @Diegis Not exactly. – egreg Sep 20 '15 at 23:20