Awk’s Authors Used Their Own Names to Demonstrate Its Search Pattern
Bell Labs designed a tool for disposable data chores. Its surviving implementation now handles Unicode and quoted CSV without surrendering the small-program approach.

The authors put themselves in the search expression. Fair enough. If you are explaining a language named after your own initials, the sample data has already volunteered.
The GNU awk history identifies the designers as Alfred V. Aho, Peter J. Weinberger and Brian W. Kernighan. It dates the original Bell Labs version to 1977. Their little naming gag sits inside a serious attempt to make small data-processing jobs less annoying.
Their paper demonstrates a regular expression matching the three surnames, with either an uppercase or lowercase initial. Its explanation says the pattern prints names “whether capitalized or not.” The brackets cover the initial letters in this example; they do not magically make the rest case-insensitive.
The other trick is what they leave out. With a pattern but no action, awk prints a matching line. The input loop is automatic, too. Here is the same idea as a runnable command:
awk '/[Aa]ho|[Ww]einberger|[Kk]ernighan/' names.txt
The vertical bars mean alternatives. A line can contain any one of the names and qualify. These are substring searches, so a surname embedded in a longer word also gets through. If you need a whole field, you have another decision to make before that command becomes your report.
The authors were targeting chores whose setup could outweigh their useful work: selecting records, rearranging fields, computing totals. A program might be needed once, yet still demand input handling and conversions. Awk took responsibility for the repetitive machinery. You could concentrate on what made a record interesting.
A one-off filter should not require the ceremony of opening a new software company. But a short command only helps when its defaults are familiar enough that the next person can read it without conducting an archaeological dig through a terminal history.
The One True Awk repository connects that old example to a maintained implementation described in the authors’ 2024 second edition of The AWK Programming Language. Its README says the newer version handles UTF-8 and CSV input. The everyday mess in the data has kept changing.
For strings, functions such as length and substr count Unicode code points. The README cautions that code points are not necessarily characters, a distinction worth remembering before promising someone that a length check measures what they see on screen.
Its --csv option also handles quoted fields containing commas or embedded newlines, with doubled quotes inside quoted fields. Simply splitting every line at each comma cannot do that job. In CSV mode, the usual FS field-separator setting is ignored. Check which awk implementation is installed before expecting that option to work.
Awk still offers a quick way to inspect a file before lunch. The useful sequel to the authors’ name joke is sitting in your own scripts: which pieces of input-handling code can the tool already do, and which assumptions about your data still need to be written down?