Path: utzoo!utgpu!news-server.csri.toronto.edu!rpi!zaphod.mps.ohio-state.edu!pacific.mps.ohio-state.edu!linac!att!att!andante!alice!dmr
From: dmr@alice.att.com (Dennis Ritchie)
Newsgroups: comp.unix.questions
Subject: Re: "Glob"
Message-ID: <20226@alice.att.com>
Date: 18 Apr 91 06:54:13 GMT
Organization: AT&T Bell Laboratories, Murray Hill NJ
Lines: 22

I am the only person who ever knew the true etymology of "glob."
Unfortunately, I have forgotten it.

I do remember, of course, that the program in early PDP-11 Unix
that expanded ?* for the shell was called /etc/glob, but the
detail that escapes is why, exactly, that name was picked.
Deep hypnotic therapy has revealed that it is cognate with
"global," and of this I am confident.  The naggingly lost detail
is the relationship of the name to the function.  I can only
guess that I reasoned that the * notation allowed commands
to apply globally to a directory.

BTW, the Jargon file is wrong in connecting /etc/glob with
the Bourne shell.  Bourne was the one who integrated file
expansion into his shell, and thereby obsoleted /etc/glob.
Also it's incorrect to say that globbing couldn't fit into
the early shells; in the 5th edition, /bin/sh was 4992 bytes,
/etc/glob 1280 bytes of program text.  Instead, the separation
was an early experiment in modularization and tool-use.

	Dennis Ritchie
	dmr@research.att.com