How to define a command for putting periods in abbreviations (initialisms)?

Question

Is there an easy way to define a command that takes a word (an initialism) and inserts a period after every character? That is, I would like to define a command \init{abc} that outputs "A.B.C." (making no assumption on the number of characters in the input argument).

The intended use is to be able to switch the output of \init{#1} to \uppercase{#1} or \textsc{#1} depending on the desired style, or to distinguish between initialism like C.P.U. (each letter pronounced separately) and NASA (pronounced as one word) like the New Yorker does.

Answer 1

You can split the input at each token and then deliver them separated by periods; with \initformat you can choose the formatting, the argument should be a one parameter macro such as \MakeUppercase or \textsc.

\documentclass{article}
\usepackage{xparse}

\ExplSyntaxOn
\NewDocumentCommand{\init}{m}
 {
  \tanh_init:n { #1 }
 }
\NewDocumentCommand{\initformat}{m}
 {
  \cs_set_eq:NN \tanh_init_format:n #1
 }

\seq_new:N \l_tanh_init_seq

\cs_new_protected:Nn \tanh_init:n
 {
  \seq_set_split:Nnn \l_tanh_init_seq { } { #1 }
  \tanh_init_format:n { \seq_use:Nn \l_tanh_init_seq { . } .\@ }
 }
\ExplSyntaxOff

\initformat{\MakeUppercase} % initialize

\begin{document}

\init{abc}

\initformat{\textsc}

\init{abc}

\end{document}

If you want to remove the trailing period when the initialism is followed by a period, you can check for it and take a decision. In the example I used \xspaceskip to make the effect more visible.

\documentclass{article}
\usepackage{xparse}

\ExplSyntaxOn
\NewDocumentCommand{\init}{m}
 {
  \tanh_init:n { #1 }
 }
\NewDocumentCommand{\initformat}{m}
 {
  \cs_set_eq:NN \tanh_init_format:n #1
 }

\seq_new:N \l_tanh_init_seq

\cs_new_protected:Nn \tanh_init:n
 {
  \seq_set_split:Nnn \l_tanh_init_seq { } { #1 }
  \tanh_init_format:n { \seq_use:Nn \l_tanh_init_seq { . } }
  \tanh_init_period:
 }
\cs_new_protected:Nn \tanh_init_period:
 {
  \peek_charcode:NTF . { \@ } { \tanh_init_format:n { . } \@ }
 }
\ExplSyntaxOff

\xspaceskip=20pt

\initformat{\MakeUppercase} % initialize

\begin{document}

\init{abc} whatever

\init{abc}. A period!

\initformat{\textsc}

\init{abc} whatever

\init{abc}. A period!

\end{document}

If you want to preserve kerning, you need to make the test about . beforehand:

\documentclass{article}
\usepackage{xparse}

\ExplSyntaxOn
\NewDocumentCommand{\init}{m}
 {
  \tanh_init:n { #1 }
 }
\NewDocumentCommand{\initformat}{m}
 {
  \cs_set_eq:NN \tanh_init_format:n #1
 }

\seq_new:N \l_tanh_init_seq

\cs_new_protected:Nn \tanh_init:n
 {
  \seq_set_split:Nnn \l_tanh_init_seq { } { #1 }
  \tanh_init_period:
 }
\cs_new_protected:Nn \tanh_init_period:
 {
  \peek_charcode_remove:NTF .
   {
    \tanh_init_format:n { \seq_use:Nn \l_tanh_init_seq { . } . } \spacefactor\sfcode`.~ 
   }
   {
    \tanh_init_format:n { \seq_use:Nn \l_tanh_init_seq { . } . } \@
   }
 }
\ExplSyntaxOff

\xspaceskip=20pt

\initformat{\MakeUppercase} % initialize

\begin{document}

\init{abp} whatever

\init{abp}. A period!

\initformat{\textsc}

\init{abp} whatever

\init{abp}. A period!

\textsc{a.b.p.} A period!

\end{document}

Answer 2

One way to do it:

\documentclass{article}

\begin{document}
\newcommand{\init}[1]{%
  \let\nextinit=\relax
  \if\relax\noexpand#1\relax\else#1.\let\nextinit=\init\fi\nextinit}

\init IBM\relax{} is an abbreviation, as is
\init GNU\relax.  What is \init SPECTRE{}?
\end{document}

\init is now defined as a function that takes a single argument. Thus, in \init IBM, first, only the I will be the argument to \init. The idea is that \init checks whether its argument equals \relax. If yes, it does nothing and stops. If not, then it prints the argument plus a ., and it replicates itself. Thus it continues, until it meets a \relax. It also stops at {} (even though that's not strictly speeking a \relax, but well).

You could also try to implement other stopping criteria, such as a special character, using \@ifnextchar, see Understanding \@ifnextchar.