cp /usr/bin/grep ack Find the needles ./ack needle haystack Find the silver needles ./ack silver.*needle Find all needles except lead ones ./ack

cp /usr/bin/grep ack Find the needles ./ack needle haystack Find the silver needles ./ack silver.*needle Find all needles except lead ones ./ack '[^^][^e.][^a.][^d.] needle' haystack That last one could be tricky if there's other types of needles with names like "ead needle" or "mead needle". But using the haystack he gives us BRE can do the job, easily. Perl regex may be easy to use but they are inferior from a performance perspective. As someone else said, they're slower than BRE or ERE. Moreover, even if speed is not an issue, you pay a price in the amount of memory you will need compared with line-based utilities like, e.g., sed and awk. Find the needles sed '/needle/!d;/needle/q' haystack Find the silver needles sed '/silver needle/!d;/silver needle/q' haystack Find all needles except lead ones sed '/lead needle/d;/needle/!d/needle/q' haystack My preference is to use (f)lex if I want a fast "parser" (scanner). Its regex is more than adequate.