I do not follow your claim that
The[re is] unnecesary space ... between $$1+1$$
and \section
.
First off, let's posit that you should never use $$
in a LaTeX document to initiate and terminate single-line display math mode. Instead, do use \[
and \]
. For more information on this subject, please see the posting Why is \[ ... \]
preferable to $$ ... $$
?
Second, please compare the amounts of vertical whitespace in the following two-column example:

The only difference is the use of \[ ... \]
on the left and of $ ... $
, i.e., of inline math mode within text mode, on the right. Observe that the vertical distance between the math expressions and the subsequent section headers is the same. This vertical distance is controlled by the specification of the \section
command, not by whether the final preceding material is typeset in text mode or display-math mode.
Addendum: If one were to modify the preceding example by loading the parskip package, there would be more whitespace at the first paragraph break, i.e., between ldkjfal;sdfkj
and $1+1=2$
. However, the vertical space between the math-y expressions and the subsequent sectioning header lines would be the same as in the example shown above. Conclusion: LaTeX does not insert "unnecessary" space between a displayed equation and a subsequent \section
directive.
\documentclass[twocolumn]{article}
\begin{document}
% left-hand column:
ldkjfal;sdfkj
\[
1+1=2
\]
\section{A new section}
\newpage % force a column break
% right-hand column:
ldkjfal;sdfkj
$1+1=2$
\section{Another new section}
\end{document}
$$...$$
should not be used in LaTeX as it does not follow the LaTeX settings. (3) Always provide a full but minimal exmaple others can copy and test as is. When we know exactly what you are doing. Here we have no idea about your setup.parskip
is not what you think. You always get a new paragraph by leaving a blank line in the TeX file. The difference is that withparskip
you get vertical space in the output instead of indentation, which is a stylistic decision. Look in the books you have and count how many of them use this style: few, if any.