Now I'm just guessing, but it seems to me that what they want is the text that most closely corresponds to a given action. The concern is the text, not the regex. Since I haven't seen the data you work with, this might be way off base, but perhaps a better way would be to find all regexes that will match with the text and then choose the text with the smallest "edit distance" against a target text?
Edit distance is generally concerned with the number of additions, deletions and substitutions necessary to transform one string into another. There are a number of CPAN modules which can calculate this for you. See Text::LevenshteinXS for a representative example.
Cheers,
Ovid
New address of my CGI Course.
In reply to Re: Most specific pattern (edit distance)
by Ovid
in thread Most specific pattern
by thor
| For: | Use: | ||
| & | & | ||
| < | < | ||
| > | > | ||
| [ | [ | ||
| ] | ] |