Relay-Version: version B 2.10 5/3/83; site utzoo.UUCP Posting-Version: version B 2.10.1 6/24/83; site ut-ngp.UUCP Path: utzoo!watmath!clyde!burl!hou3c!hocda!houxm!houxz!vax135!floyd!cmcl2!seismo!ut-sally!ut-ngp!kjm From: kjm@ut-ngp.UUCP Newsgroups: net.lang.c Subject: Re: Algol-style vs C-style semicolons Message-ID: <679@ut-ngp.UUCP> Date: Fri, 8-Jun-84 17:44:49 EDT Article-I.D.: ut-ngp.679 Posted: Fri Jun 8 17:44:49 1984 Date-Received: Sun, 10-Jun-84 00:15:38 EDT References: <940@dciem.UUCP> <3000028@uokvax.UUCP> Organization: Comp. Center, Univ. of Texas at Austin Lines: 20 > ... Look at C's excuse for a grammar, and you will see massive > amounts of extranea in the production for "statement" whose sole > purpose in life is to make it possible to avoid having to type a > semicolon after the } in a compound statement. C therefore does > *NOT* use semicolon as a statement terminator, no matter how many > times you see people claim it does. > > I think that Huffman coding of tokens is not the best principle > for language design. > > James Jones OK, so C effectively has two kinds of statement terminator: ';' and '}'. So what? The two tokens terminate two different kinds of objects. BTW, if C had been done with Huffman coding, wouldn't reserved words be only one character long? :-) Ken "Smurf-shredder" Montgomery ...{ihnp4,seismo,ctvax}!ut-sally!ut-ngp!kjm