Path: utzoo!mnetor!uunet!kddlab!icot32!icot!chik
From: chik@icot.JUNET (Chikayama Takashi)
Newsgroups: comp.lang.prolog
Subject: Re: BSI standards
Message-ID: <796@icot.JUNET>
Date: 11 Mar 88 08:07:19 GMT
References: <8803082357.AA01587@decwrl.dec.com> <5334@utah-cs.UUCP> <751@cresswell.quintus.UUCP>
Reply-To: chikayama@icot.JUNET (Chikayama Takashi)
Organization: ICOT, Tokyo, Japan
Lines: 13

In article <751@cresswell.quintus.UUCP> ok@quintus.UUCP (Richard A. O'Keefe) writes:
>What problems are people having that could be reduced by having a separate
>character type (as opposed to having characters coded as integers but with
>a tolerable notation such as `x`)?

I had at least one such experience.  When I wrote a CROSS compiler (or
rather, a cross preprocessor, that translates some language with
Prolog-like syntax to another) several years ago, it was quite nasty
that 0'x could not be distinguished from 120 once read in.  On the
target machine that uses different character codes, 120 should mean
120 but 0'x shouldn't.

Takashi Chikayama