Count the number of words on each beamer slide?

Question

Is there a way to automatically count the number of words on each beamer slide, and to display it on the slide? I would like this feature to be available only when I am creating draft versions of the slides. Ideally, I should be able to toggle this feature off when I create the final version of the slides for distribution.

MWE

\documentclass{beamer}

\begin{document}

\begin{frame}{First Slide}
This is my first slide.

% I want to generate the word count below automatically
nwords = 7
\end{frame}

\begin{frame}{Second Slide}
This is my second slide.
It has more words than the first slide.

% I want to generate the word count below automatically
nwords = 15
\end{frame}

\end{document}

The reason why I ask this question is because some people recommend that a slide should not have too many words. For example, an article by Presentation Magazine titled How many words should I have on each slide? gives the rule of thumb of "less than 40 words per slide." It would thus be useful to know how many words I have in each slide.

Answer 1

Good thing, that the beamer frame collects its body in a token list like the amsmath environments. Thus we can simply examine the token list where the contents are stored. However, we have to hack the beamer frame for that. All the arguments to the frame, in this case {First Slide}, are counted as well. Brace groups count as one token and other oddities apply, see the documentation of \seq_set_split.

\documentclass{beamer}
\usepackage{xparse}

\ExplSyntaxOn

\cs_generate_variant:Nn \seq_set_split:Nnn { Nno }

\NewDocumentCommand \setwordcount { m m }
{
  % split the frame contents at spaces
  \seq_set_split:Nno \l_tmpa_seq { ~ } { \the#2 }
  % remove all mentions of the counter variable
  \seq_remove_all:Nn \l_tmpa_seq { #1 }
  % set counter
  \tl_gset:Nx #1 { \seq_count:N \l_tmpa_seq }
}

\ExplSyntaxOff

\makeatletter

\def\beamer@frameenv{%
  \def\beamer@process@envbody{\endgroup%
    \setwordcount\insertwordcount\beamer@envbody% Count number of space delimited tokens in the input
    \expandafter\expandafter\expandafter\beamer@framecommand\expandafter\beamer@frameoptions\expandafter{\the\beamer@envbody}}%
  \global\beamer@envbody{}\def\beamer@begin@stack{b}%
  \begingroup
  \let\frame\beamer@collect@@body
  \def\beamer@process@envbody{\frame}%
  \beamer@process@envbody%
}

\makeatother

\begin{document}

\begin{frame}{First Slide}
  This is my first slide.
  \insertwordcount
\end{frame}

\end{document}

---

It may be better to count words with l3regex ... but maybe not. Who knows? This is recursive in subgroups and doesn't count control sequences but math material is still tricky.

This produces similar output but with count 7.

\documentclass{beamer}
\usepackage{xparse}

\ExplSyntaxOn

\cs_generate_variant:Nn \regex_count:nnN { noN }

\NewDocumentCommand \setwordcount { m m }
{
  % regex match words
  \regex_count:noN { \w+ } { \the#2 } \l_tmpa_int
  % set counter
  \tl_gset:Nx #1 { \int_to_arabic:n \l_tmpa_int }
}

\ExplSyntaxOff

\makeatletter

\def\beamer@frameenv{%
  \def\beamer@process@envbody{\endgroup%
    \setwordcount\insertwordcount\beamer@envbody% Count number of space delimited tokens in the input
    \expandafter\expandafter\expandafter\beamer@framecommand\expandafter\beamer@frameoptions\expandafter{\the\beamer@envbody}}%
  \global\beamer@envbody{}\def\beamer@begin@stack{b}%
  \begingroup
  \let\frame\beamer@collect@@body
  \def\beamer@process@envbody{\frame}%
  \beamer@process@envbody%
}

\makeatother

\begin{document}

\begin{frame}{First Slide}
  This is my first slide.
  \insertwordcount
\end{frame}

\end{document}