Path: utzoo!utgpu!water!watmath!clyde!att!osu-cis!killer!ames!lll-tis!lll-winken!uunet!mcvax!ukc!reading!onion!cf-cm!cybaswan!iiit-sh
From: iiit-sh@cybaswan.UUCP (s.hosgood)
Newsgroups: comp.sys.ibm.pc
Subject: Re: US PC programmers still live in a 7-bit world!
Summary: Ascii gives problems even in Britain..
Message-ID: <29@cybaswan.UUCP>
Date: 8 Jul 88 21:14:58 GMT
References: <1988Jun22.223158.1366@LTH.Se> <699@omen.UUCP> <1288@odyssee.UUCP>
Lines: 55

The world needs  an 8-bit international character set.

The 'national characters' in ascii give problems to us Brits too, though we've
got the 'right' alphabetic characters, we run into the fact that the
pound sign often appears on our terminals instead of the dollar sign, or
sometimes (rarely) instead of the hash mark. The nuisance is that the treatment
is inconsistent, and frequently printers on a system disagree with half the
VDUs! Sometimes you type a document with a price in pounds on one machine only
to get it in dollars by magic in the final output. The conversion is done with
no reference to the current exchange rate of course!

Hex constants in assembler etc look silly if the leading dollar gets printed
as a pound. 'Include' lines in C sometimes come out as '(pound) include'. The
problems are so silly that no-ones come up with a systematic cure. You can get
used to anything, but should you have to?

I look forward to the day 8-bit codes take over. Everyone will have the full
set of codes, and many more like the 'degrees' symbol and Greek letters etc
etc. However, even if you allow for the letters of French, German, Icelandic
and many others, somewhere, someone will miss out. The Celtic languages often
need vowels with circumflexes. Esperanto needs quite a few other letters
with circumflexes, and the 'u' needs an upside-down one! You would need  the
letters in upper and lower case too..

To be fair to everyone (who uses the Roman or Greek characters), when a
newer character set is accepted, it ought to have the 'more popular' characters
built in (i.e. for the use of the bigger natural languages), but surely it
ought also to enshrine a standard way for the slightly off-beat characters
to be sent as a digraph??

If you go and read the esperanto group, most of the arguments at the moment
are about the various ways of representing thier circumflex characters. Just
about everyone does it differently. They all involve digrams however, which
are hard to read.

P.S.
Can you reprogram the character set in an IBM-PC to put in the characters you
want?? Rumour has it that INT $1F points to a table of characters with codes
$80 - $FF. It seems to be ignored on my clone-XT (monochrome adaptor fitted).
I want to show those Esperanto digrams in a readable way!


----------------------------------------------------------------------


                       Steve Hosgood BSc,
             Image Processing and Systems Engineer,
         Institute for Industrial Information Techology,
             Innovation Centre, University of Wales,
                         Swansea SA2 8PP

    Phone (0792) 295213		JANET: iiit-sh@uk.ac.pyr.swan
    Fax (0792) 295532
    Telex 48149

       My views are not necessarily those of my employers!