Re: HTML - i18n / NCR & charsets

Peter Flynn (pflynn@curia.ucc.ie)
30 Nov 1996 13:27:34 +0000 (GMT)


Date: 30 Nov 1996 13:27:34 +0000 (GMT)
From: Peter Flynn <pflynn@curia.ucc.ie>
Subject: Re: HTML - i18n / NCR & charsets
In-reply-to: <199611300702.XAA15425@sweden.it.earthlink.net> (message from
To: davidp@earthlink.net
Cc: www-html@w3.org, www-international@w3.org, unicode@unicode.org
Message-id: <199611301327.NAA02320@curia.ucc.ie>

   ISO8879 names for Windows CP 1252 80-9F (128-160) entities:

   83 (131) --   ?   -- florin

What's a florin? I know it's the old UK name for what was two
shillings, but Bill obviously means something else here.

   8B (139) --   ?   -- guilsinglleft
               &laquo;

   9B (155) --    ?   -- guilsinglright
               &raquo;

   9F (159) -- &Yuml; -- Ydieresis

Y diaeresis is a non-existent character, according to the experts on
TYPO-L, who have just discussed this in depth. It was included in both
ISOlat1 (lc) and ISOlat2 (uc) as well as the IBM pc character sets in
the mistaken belief that it actually existed in some language. It was
in fact transcribed in error, either from an &ijlig; or something
similar by whoever was representing the character sets to Geneva at
the time, and no-one was prepared to bite the bullet and say "this
does not exist", for fear of being proved wrong, and thus attacked for
failing to cater for whatever language was supposed to require
it. Various claims have been made for its existence in Dutch, Turkish,
and other less populous languages, but none of these have been
demonstrated. It is possible that some orthographer has found a use
for it in transcribing a spoken language, in which case it must
obviously be kept, but the consensus of the typographers is that it is
redundant (and this included a lot of people with very significant
foreign-language typesetting experience). The logs of the discussion
can be retrieved in the usual manner from listserv@danann.hea.ie by
sending an email containing the command GET TYPO-L LOGyymm 
(where yy=96,and mm={08|09|10|11} -- I _think_ the discussion started
in August, possibly September).

///Peter