Path: utzoo!utgpu!news-server.csri.toronto.edu!rutgers!cs.utexas.edu!samsung!munnari.oz.au!goanna!ok
From: ok@goanna.cs.rmit.oz.au (Richard A. O'Keefe)
Newsgroups: comp.lang.c
Subject: Re: Internationalisation
Message-ID: <3641@goanna.cs.rmit.oz.au>
Date: 30 Aug 90 04:51:42 GMT
References: <24141@megaron.cs.arizona.edu> <134@blekko.UUCP> <1911@islay.tcom.stc.co.uk>
Organization: Comp Sci, RMIT, Melbourne, Australia
Lines: 44

In article <1911@islay.tcom.stc.co.uk>, rmj@tcom.stc.co.uk (Rhodri James) writes:
> Mind you, arguing that "this is the way System V does it, so get used to
> it" nearly lost you my sympathy.

It wasn't *supposed* to *keep* your sympathy.
There are a lot of things in System V Release 4 I don't particularly
care for (having gone to the trouble of learning how X/Open handles
internationalisation, I didn't really appreciate discovering that the
new Official way of doing it was different, and while the TLI routines
may perhaps be a considerable improvement on sockets, I have yet to
find anything which explains them clearly enough for me to use them).

The point I was making is that *customers* are going to expect SVR4
programs to behave in a particular way.  SVR4 has a convention for
generating multi-line error messages (SVR4 is an adventure; if you
win you find you're playing VMS), and it has lots of features for "locale"
support (if you want the "C" word rather than the "UNIX" word, which
has been current for, oh, at least 3 years).  In a couple of years
time, customers are going to expect programs to follow the UNIX Way,
just as Macintosh customers expect Mac programs to follow the Mac Way.
So we had better get used to it if we want to produce programs that
the next decade's UNIX customers will continue to be willing to buy.

By the way, "language switching" would NOT be an appropriate replacement
for the word "internationalisation" because the latter covers rather more.
Wales and the USA both use English (Wales also uses Welsh and the US also
uses Spanish).  But they don't represent dates the same way, and they
don't use the same symbols for currency.  Internationalisation refers to
collating order, date and time representation, currency representation,
and a couple of other things I forget as well as the language that messages
are displayed in.  A program portable between different locales will _not_,
for example, assume that everyone has a three-part name, a common US-ism.

> How Unix of any sort has become the
> dominant operating system is beyond me, it's not as if it's actually
> very good or anything :-\

"Democracy is the worst possible political system, except for all the others."
Unix hasn't succeeded by being particularly good, but by not being
excruciatingly bad (unlike xx-xxx, xx/xxx, xxx-xx, xxx, or xxx -- names
changed to protect _me_).

-- 
You can lie with statistics ... but not to a statistician.