W3C home > Mailing lists > Public > ietf-charsets@w3.org > April to June 2000

Registration of new charset SCSU

From: Markus Scherer <markus.scherer@jtcsv.com>
Date: Thu, 15 Jun 2000 08:41:26 -0700
To: charsets <ietf-charsets@iana.org>
Message-id: <3948F926.204890DC@jtcsv.com>
(This is an updated proposal that incorporates feedback from
Kenneth Whistler, Harald Tveit Alvestrand, and Martin J. Dürst.)

Charset name: SCSU

Charset aliases: (none, except for the implicit csSCSU)

Suitability for use in MIME text: No

Published Specification:
    Unicode Technical Report #6
    "A Standard Compression Scheme for Unicode"
    http://www.unicode.org/unicode/reports/tr6/

    CCS & CES: The SCSU charset is a combination of the
    Unicode and ISO 10646 Coded Character Set (CCS) with
    the Character Encoding Scheme (CES) specified in
    the above document. It covers exactly the
    UTF-16-reachable subset of ISO 10646.
    SCSU can also be classified as a Transfer Encoding
    Syntax (TES), but one specifically designed for
    Unicode/ISO 10646.

ISO 10646 equivalency table: Same as specification.

Additional Information:
    SCSU is an encoding (CES/TES) of Unicode/ISO 10646
    that allows significant size reduction of text compared
    to UCS Transformation formats. It approximates the size of
    text that is otherwise achieved with language-specific
    charsets while encoding all of Unicode (up to U-0010ffff).
    SCSU is byte-based, which helps further, traditional
    compression (LZW etc.).
    It is stateful and uses all byte values including NUL.
    CRLF may or may not be represented by 0x0d 0x0a depending
    on the encoder and the text.
    Encoders can be trivial by emitting one command byte (0x0f)
    followed by the text in UTF-16BE. Fairly simple encoders
    yield good results with average text of any length.
    Decoding is simple and requires no mapping tables.
    If no control characters other than NUL, TAB, CR, and LF
    are used, then text in US-ASCII or ISO-8859-1 is already
    valid SCSU text.
    There is a Unicode signature byte sequence defined
    (0e fe ff, see specification).

    SCSU is of no use for applications that require a canonical
    representation of text. It is not suitable for
    process-internal use.
    This is an intentional part of its design.

Personal & email address to contact for further information:
    Markus W. Scherer
    IBM Java Technology Center
    10275 N. DeAnza Blvd
    Cupertino, CA 95014-2237

    markus.scherer@jtcsv.com
    schererm@us.ibm.com

Intended usage: LIMITED USE


Thank you for your consideration,

markus
Received on Thursday, 15 June 2000 11:47:13 GMT

This archive was generated by hypermail 2.2.0+W3C-0.50 : Monday, 5 June 2006 15:10:51 GMT