Path: utzoo!utgpu!water!watmath!clyde!att!osu-cis!tut.cis.ohio-state.edu!rutgers!cmcl2!husc6!ogccse!metheus!todd
From: todd@metheus.UUCP (Todd Stewart)
Newsgroups: comp.sys.ibm.pc
Subject: Re: Microsoft Vs. Borland
Message-ID: <141@metheus.UUCP>
Date: 24 Oct 88 20:05:00 GMT
References: <876@galaxy> <213400010@s.cs.uiuc.edu> <1709@dataio.Data-IO.COM>
Reply-To: todd@metheus.UUCP (Todd Stewart)
Organization: Metheus Corporation, Portland Or.
Lines: 18

In article <1709@dataio.Data-IO.COM> bright@dataio.Data-IO.COM (Walter Bright) writes:
>... It also has some pointless typedefs such as:
>	typedef unsigned char u_char;
>Really, what's that worth?

Mostly it takes the sting out of saying what you really mean.
"unisigned char" is so unwieldy that most go ahead and use default signed
data types even when handling unsigned data e.g. array indices, etc.

These subtle bugs go undetected by the original authors, but break in nasty
and hard to debug ways when the code is later ported to a different configura-
tion or host.

I'm all for any typedef convention that encourages accurate programming by
making it reasonable to say what you really mean.

	Todd Stewart
	Metheus Corporation