I believe that option p
of figure
environment is necessary (if you don't want that big floats filling up almost an entire page are placed at the end of document, followed by smaller floats which could otherwise easily mix with text) but is unsatisfactory in its default behaviour. As the following MWE shows (but many other examples could be made), with the p
option of figure
(image on the left) the second figure ends up on a float-only page, whereas without the p
option (image on the right) the second figure, correctly IMHO, mix with the text.
\documentclass{article}
\usepackage{mwe}
\begin{document}
\lipsum
\begin{figure}[htbp]
\centering
\includegraphics[width=.8\textwidth]{example-image-1x1}
\caption{A figure caption.}
\end{figure}
\begin{figure}[htbp]
\centering
\includegraphics[width=.8\textwidth]{example-image-1x1}
\caption{A figure caption.}
\end{figure}
\lipsum
\end{document}
Is it possible to adjust the utilization of float-only pages so that they are used only when really needed? In my opinion the criterion triggering the employment of float-only pages should be something like "use float-only pages only when more than 85% (or so) of available text height gets filled up".
Thanks in advance for any clue.
P.S. I know that one could use the p
option only "when needed", but if you have dozens of images in each one of dozens of documents you really do not want to check every single image for the p
option.
\includegraphics
does not have ap
option (or any positioning logic at all you meanfigure
just increase float page fraction, I would guess.