I always like to keep everything in plain text but sometimes it is a pain in the ass the concept that content and presentation is mixed in these files. If you keep lines wrapped to be under some column width, that implies that using a regular expression to locate some group of words is gonna be problematic, as regexp also have the pervasive concept of line.
That reminds me that I should read more about structural regular expressions which precisely try to fix this and other issues.
So Vim to the rescue it is. GNU grep features
--perl-regexp but the doc states that it is experimental. On Vim, we have the pattern
\_ which extends any character set with the new-line. Thus with the whitespace character set
\s we can search for a contiguos pair of words which could occur in different consecutive lines:
“I define UNIX as 30 definitions of regular expressions living under one roof.” —Donald Knuth