Relay-Version: version B 2.10 5/3/83; site utzoo.UUCP
Path: utzoo!watmath!clyde!akgua!mcnc!decvax!harpo!ihnp4!zehntel!hplabs!sri-unix!O'KeefeHPS
From: O'KeefeHPS@sri-unix.UUCP
Newsgroups: net.lang.prolog
Subject: A coding trick for grammar rules.
Message-ID: <12452@sri-arpa.UUCP>
Date: Wed, 4-Apr-84 06:05:13 EST
Article-I.D.: sri-arpa.12452
Posted: Wed Apr  4 06:05:13 1984
Date-Received: Mon, 23-Apr-84 01:45:15 EST
Lines: 31

From:  O'Keefe HPS (on ERCC DEC-10)

Someone here today wanted to write a grammar rule which
could detect the end of a word.  It isn't always possible
to check the next character, becuase the word might be
at the end of the string.  What was needed was something
like

end_of_word --> ( end_of_string | [C], {not_in_word(C)} ).

It finally dawned on me that I could write

        end_of_string(X, X).

This dodge may be useful to other hackers.  REAL hackers
will doubtless prefer to write

end_of_word --> ( = | [C], {not_in_word(C)} ).

If one prefers not to eat the character following the word,
one can write

end_of_word --> = .     % space needed for tokeniser!
end_of_word, [C] --> [C], {not_in_word(C)}.

Of course the clean way would have been to match the longest
possible string of letters, digits, and underscores, but we
were after a quick patch.
I'm not at all sure that I care for any of this...

Possibly we could start a Prolog Joke Book ?