Relay-Version: version B 2.10 5/3/83; site utzoo.UUCP
Posting-Version: version B 2.10 5/3/83; site houxq.UUCP
Path: utzoo!linus!decvax!harpo!gummo!whuxlb!pyuxll!eisx!npoiv!npois!hogpc!houxq!eli
From: eli@houxq.UUCP
Newsgroups: net.lang
Subject: Does C have I/O, or doesn't it?
Message-ID: <425@houxq.UUCP>
Date: Tue, 30-Aug-83 23:58:04 EDT
Article-I.D.: houxq.425
Posted: Tue Aug 30 23:58:04 1983
Date-Received: Thu, 1-Sep-83 01:34:51 EDT
Organization: American Bell, Holmdel NJ
Lines: 23

Michael Condict (...!cmcl2!csd1!condict) posted a news article arguing
that, in spite of any language definitions or what others may say,
the C language does in fact have I/O.  So far I have seen 3 replies
telling Mr. Condict what he obviously already knows: the C language
definition does not make any reference to I/O, and that I/O is done
using function calls to a support library.

Those people replying to the article with statements to the effect
indicated above seem to have missed a major point of the discussion: How does
the inclusion of I/O affects the "goodness" of the language, and specifically,
how does it affect portability of programs written in the language?

In any case, I think the basic problem with these replies is that the word
"language" is being used to mean the specification of what the compiler
must parse, while Mr. Condict is using it to mean what you must code
in your program to tell the computer to produce useful results.

If we can get our terminology straight, we may find more interesting things
to disagree on.

					Eli Mantel (houxq!eli)
					AT&T Information Systems Laboratory
					Holmdel, New Jersey