# Is there a parenthesis code that work in all contexts and arguments?

This question is about having the most generic parenthesis code if one doesn't know the contents size of the parenthesis a priori.

If I want to write a function or parenthesized equation I write, for example, $f(...)$.

If I want to use the same code regardless of the contents of the parenthesis ... one tends to find that \left(...\right) is likely to look better if the contents are big.

However the spacing is not correct, even for small contents. In the sense that for example f(x) is not visually equivalent to f\left(x\right). So to cover for that case I write f\!\left(...\right).

I recently found that this works in many contexts but not in a superscript. For example $e^{i\arg\!\left(z\right)}$, produced a badly spaced formula:

Is there a parenthesis code that works more or less well regardless of content and the context?

It looks like \! is pretty close to a solution, so preferably without using additional packages. (references to packages are still welcome).

EDIT: From @Manuel's answer below, it looks like the most elegant solution is to use the package mleftright and leave all the \left(, \right) commands as they are. It also looks like one can use \mathopen{}\mathclose{\left(...\right)} if one doesn't want to include packages.

Here it is a sample to show that all solutions give a consistent output:

\documentclass[]{article}
\begin{document}

$i \arg(z)$
$i \arg\left(z\right)$
$e^{i \arg\!\left(z\right)}$
$e^{i \arg\mathopen{}\mathclose{\left( z\right)}}$
$i \arg\mathopen{}\mathclose{\left( z\right)}$
$i \arg\mathopen{}\mathclose{\left(\frac{z}{y}\right)}$

\end{document}


\documentclass[]{article}
\usepackage{mathtools,mleftright}
\mleftright
\begin{document}

$i \arg(z)$
$i \arg\left(z\right)$
$e^{i \arg\!\left(z\right)}$
$e^{i \arg\mathopen{}\mathclose{\left( z\right)}}$
$i \arg\mathopen{}\mathclose{\left( z\right)}$
$i \arg\mathopen{}\mathclose{\left(\frac{z}{y}\right)}$

\end{document}


• mleftright and \mleft and \mright? – Manuel May 25 '15 at 19:36
• Even if you fix the horizontal spacing matter -- using, say, the suggestion of @Manuel to employ \mleft and \mright -- you still run into the issue that you'll sometimes have parentheses that are too large (from a typographer's point of view) or, in other circumstances, too small. Thus, I don't think it can be said that there's automated code for parentheses that works under circumstances. – Mico May 25 '15 at 20:12
• AFAIR nath did an OK job. But it is incompatible with a lot of stuff and have redefine most math. So I agree with Mico, it takes experience to know what is best. In general, don't use autoscaling too much it often ends up overpowering an expression or (if used within normal text) disturbs line spacing – daleif May 25 '15 at 20:43
• Ah, and if you end up not using autoscaling (e.g. \bigl(, \Bigr), etc.) you should patch the \bigx commands because, by default they only use the \displaystyle font. Shameless plug. – Manuel May 25 '15 at 20:49
• For more information on why using \left and \right everywhere is not a good idea in general, check out the postings "(" or "\left(" parentheses? and Is it ever bad to use \left and \right? – Mico May 26 '15 at 3:48

\documentclass{scrartcl}

\usepackage{mathtools,mleftright}
\mleftright % Or comment this line and use \mleft and \mright instead

\begin{document}
$f\left(x\right)$
$\arg\left(z\right)$
$e^{i\arg\left(z\right)}$
\end{document}

• Do you know what is the secret of mleftright? it seems that smartly chooses the spacing before the opening (...). The size it self seems to be that of the default \left( \right). – alfC May 26 '15 at 6:42
• Basically defines, for instance \mleft( to be \mathopen{}\mathclose{\left( and \mright) to \right)}, so it ends being a \mathopen{}\mathclose{..} rather than \mathinner{..} (with \mleftright you redefine the \left/\right commands, so there's no need to change anything at all in the document. – Manuel May 26 '15 at 11:16
• from you comment I understand that \mathopen{}\mathclose{ ... \right)} is the poor's man version (or equivalent to) of mleftright. Indeed it works pretty well. – alfC May 27 '15 at 21:06
• You could say that. But I don't see the advantage of using \mathopen{}\mathclose{\left( .. \right)}. – Manuel May 27 '15 at 23:08