W3C home > Mailing lists > Public > www-international@w3.org > January to March 2008

Re: ISO 8859-1 C1 set in RFC 2157

From: Uma Umamaheswaran <umavs@ca.ibm.com>
Date: Tue, 25 Mar 2008 14:48:38 -0400
To: www-international@w3.org
Message-ID: <OFED8CF204.61CCABE5-ON85257417.0065E4E2-85257417.0067546B@ca.ibm.com>

Another point you may want to keep in mind as background ...

When one takes EBCDIC based (for ex: IBM 037) Latin-1 sets and when one
maps to 8859-1 for G sets (no loss) and C0 and C1 spaces for preserving the
EBCDIC controls ... one of these being NEL .. while flowing through IETF /
MIME networks and back into EBCDIC based systems .. it provides a nice pass
through capability for the corresponding number of control code positions.
While this is possible in a more or less transparent manner with so called
ISO-8bit encodings, one cannot do so with Windows(such as cp125x) or (PC
DOS such as cp850) based encodings without overlaying the controls into the
C1 space graphics or mapping the equivalents of C1 controls into/out of
sort of two byte sequences !!  I know the world at large may not want to
hear anything about EBCDIC !!!

One of the motivations for the C1 default in Unicode was just to leave
these alone as defaults from ISO 6429 to allow some level of pass through
capability with EBCDIC encoded data as well.

>From structure point of view windows 12xx, IBM PC DOS etc. could be
considered as modified ISO-8 structure (a la 4873 level 1), in the sense
that it has a C0, G0 and 128 slots G1 !!

Best regards, Uma
Received on Tuesday, 25 March 2008 18:49:25 GMT

This archive was generated by hypermail 2.2.0+W3C-0.50 : Tuesday, 2 June 2009 19:17:16 GMT