Path: utzoo!mnetor!uunet!husc6!cmcl2!brl-adm!umd5!ames!ucbcad!ucbvax!cbosgd!osu-cis!tut!lvc
From: lvc@tut.cis.ohio-state.edu (Lawrence V. Cipriani)
Newsgroups: comp.lang.c
Subject: Re: Language-induced errors
Message-ID: <4349@tut.cis.ohio-state.edu>
Date: 13 Jan 88 01:18:18 GMT
References: <4668@pyr.gatech.EDU> <495@xyzzy.UUCP> <9930@mimsy.UUCP> <18515@bu-cs.BU.EDU>
Organization: Ohio State Computer & Info Science
Lines: 38
Summary: mainly typos not bad programmers

In article <18515@bu-cs.BU.EDU>, hen@bu-cs.BU.EDU (Bill Henneman) writes:
> In article <609@PT.CS.CMU.EDU>, edw@IUS1.CS.CMU.EDU (Eddie Wyatt) writes
	paragraph by Eddie Wyatt on programming environments deleted
> 
> Are there programmers who really have trouble with this?  I know a
> fair number of C programmers (i.e., people who are paid to produce
> code), and none of us find these kinds of things to be sources of
> error.  Those of us with the most experience with other languages took
> the longest for that to stop being a source of error - about half a
> day.  By the time we had run the exercises in Chapter 3 of K&R, this
> sort of error was about as likely as my saying *DO* or *LOOP* when I
> mean *FOR*.
> 
> Is my experience unique?  Is there something in the air in Boston that
> makes us immune to this sort of error?
> 
> 			Skeptical in Kenmore Square

Must be the air in Boston :-).  I don't think there are many programmers
that really have trouble with this (for long), its just when you have been
typing at a keyboard for a long time it becomes easier to slip up. I don't
believe anyone is immune from this, except maybe Dennis Ritchie :-).  How
are you certain that these mistakes weren't made?

How should a language be designed so that these errors are hard
to make (would be an interesting thesis topic)?  I'd rather have the
power without the syntactic seat belts and have a syntax checker to
verify the code.

Other common "typos" in C are extra semicolons after if while for and else,
dangling elses (many programming languages have this problem), missing
parenthesis in expressions resulting in fouled up precedence, using
a bitwise or when a logical or was intended, ditto for bitwise and,
nesting comments, etc etc.  I mentioned in another posting a syntax
checker called inspect that I wrote in another article that checked
for this sort of stuff in C and C++.  I ran it on many different versions
of the UNIX operating system, and found bugs in every one of them.
There weren't a lot of bugs, though some of them were major problems.