Splitting Formula over several lines (automatically)

Question

I'd like to let LaTeX split a formula that's inside a fixed-width table. The formula is a group theoretical word (something like $a_1b_3a_5b_3^{-1}$).

It should automatically split it somewhere between the elements so that it fits in the table. I thought seqsplit might do the trick, but in the following code example:

\documentclass{article}
\usepackage{seqsplit}


\begin{document}

\begin{tabular}{p{4cm}}
 \seqsplit{$a_1b_3a_5b_3^{-1}$}\\
 \seqsplit{$a_1$ $b_3$ $a_5$ $b_3^{-1}$}\\
 \seqsplit{abcd}
\end{tabular}

\end{document}

Both variants with the formula just raise lots of errors of the form Missing { inserted \seqsplit{$a_1b_3a_5b_3^{-1}$}, whereas the abcd-sample works.

The second example actually wouldn't give the desired result anyway, because there should be no spaces between the elements.

Does anyone have a solution to this?

Answer 1

\documentclass{article}
\usepackage{array}


\begin{document}

\begin{tabular}{>{\spaceskip=1sp\relax\raggedright\arraybackslash}p{2cm}}
\hline
 $a_1$ $b_3$ $a_5$ $b_3^{-1}$ $a_1$ $b_3$ $a_5$ $b_3^{-1}$\\
\hline
\end{tabular}

\end{document}

Answer 2

You can input the long word in a single sequence, and add \hfil\allowbreak\hfilneg before each letter.

\documentclass{article}
\usepackage{xparse}

\ExplSyntaxOn
\NewDocumentCommand{\longword}{m}
 {
  \magula_longword:n { #1 }
 }

\tl_new:N \l_magula_longword_tl

\cs_new_protected:Nn \magula_longword:n
  \tl_set:Nn \l_magula_longword_tl { #1 }
  \regex_replace_all:nnN
   { ([a-z]) }
   { \c{hfil}\c{allowbreak}\c{hfilneg} \1 }
   \l_magula_longword_tl
  $\l_magula_longword_tl$
 }
\ExplSyntaxOff

\begin{document}

\begin{tabular}{p{4cm}}
 \longword{
   a_1b_3a_5b_3^{-1}a_1b_3a_5b_3^{-1}abcd
   a_1b_3a_5b_3^{-1}a_1b_3a_5b_3^{-1}
   a_1b_3a_5b_3^{-1}a_1b_3a_5b_3^{-1}
   a_1b_3a_5b_3^{-1}a_1b_3a_5b_3^{-1}
   a_1b_3a_5b_3^{-1}a_1b_3a_5b_3^{-1}abcd
 }
\end{tabular}

\end{document}