W3C home > Mailing lists > Public > www-archive@w3.org > January 2002

Re: codes:chars is many:one? [was: Undeliverable...]

From: Liam Quin <liam@w3.org>
Date: Wed, 30 Jan 2002 16:16:53 -0500
To: "Alan G. Isaac" <aisaac@american.edu>
Cc: Dan Connolly <connolly@w3.org>, "Martin J. Duerst" <duerst@w3.org>, Public archive <www-archive@w3.org>
Message-ID: <20020130161653.E5239@tux.w3.org>
oops, I forgot to send this, sorry.

On Fri, Jan 25, 2002 at 05:47:02PM -0600, Dan Connolly wrote:
> Lee, can you substantiate the claim that
> 
> " there are some coded character sets that assign two different numbers
> to the same character[Lee],"
> 
> please?

Hmm, not easily, although the Macintosh has the Euro in two or
possivbly even three positions in OS/X (128 and/or currency sign, plus
the Unicode position), I've been told recently.

I can do some research if it's important -- I don't remember the
context in which I wrote that.  Certainly I have used systems in
which the same character appeared in multiple positions, though --
what cmes to mind most readily is old hardware, such as terminals
and printers, e.g. in the UK before US ASCII and Latin 1 took hold
it was not unknown to find # or ` in more than one position.
Sometimes there was a physical switch that would move them or
would replace one of them with a  (pound sterling) sign.

Also note that PostScript Type 1 fonts (which conflate abstract characters
and codepoints, in that the mapping is code point to glyph name) often
have the same chanracter (sometimes even the same glyph) appearing
in multiple positions in the encoding.

The important point in Dan's paper, though, is that it's conceptually
possible, I think.

Best,

Liam
-- 
Liam Quin, W3C XMl Activity Lead, liam@w3.org
Received on Wednesday, 30 January 2002 16:16:59 GMT

This archive was generated by hypermail 2.2.0+W3C-0.50 : Wednesday, 7 November 2012 14:17:16 GMT