W3C home > Mailing lists > Public > www-international@w3.org > January to March 1997

Re: Translated IUC10 Web pages: Experimental Results

From: Alain LaBont/e'/ <alb@sct.gouv.qc.ca>
Date: Wed, 5 Feb 1997 11:31:25 -0500
Message-Id: <9702051631.AA29250@socrate.riq.qc.ca>
To: iso10646@listproc.hcf.jhu.edu, Unicore <unicore@unicode.org>, Unicode <unicode@unicode.org>, www-international <www-international@w3.org>, HTTP WG <http-wg@cuckoo.hpl.hp.com>, Search <search@mccmedia.com>, ISO10646 <iso10646@listproc.hcf.jhu.edu>
At 09:43 97-02-05 -0500, Misha Wolf wrote:
>Chris Pratley wrote:
>
>>[snip]
>
>>Our assumption was that UTF-8 was the only Web-safe encoding that was
>>reasonably likely to be adopted by browsers in the near future. Is that
>>the consensus, or are raw UCS2 encodings being considered actively by
>>people on this alias? 
>
>I think it very unlikely that plain 16-bit Unicode will be adopted by 
>browsers in the next year or two.  The two encoding schemes which will 
>be widely used to encode Unicode Web pages are:
>
>   1.  UTF-8 (see <http://www.reuters.com/unicode/iuc10/x-utf8.html>).
>   2.  Numeric Character References (see
<http://www.reuters.com/unicode/iuc10/x-ncr.html>).
>
>The second scheme is intriguing as it does not require the use of any 
>octets over 127 decimal (7F hex).  Accordingly, it is legal to to label 
>such a file as, eg, US-ASCII, ISO-8859-1, X-SJIS, or any other "charset" 
>which has ASCII as a subset.  Browser vendors: Please check your products 
>against the pages referenced above.
>
>>[snip]
>
>Regards,
>Misha


I do not understand why it is more complicated to use UCS-2 than any other
scheme (apart from the little-endian problem, which should be deprecated in
the state of the art of the XXIth Century, it is a patch!) The web requires
8- bits-per-octet encoding (thank God! otherwise even UTF-8 would not work)
as its default character set is ISO/IEC 8859-1. A wise implementer should
implement at least:

        -Latin 1
        -UTF-8
        -entity names
        -UCS-2 (big-endian at least, little-endian as a patch if indicated
                clearly!)

Anyway the logic, one the source data has been normalized, should be the
same after all. I am pretty sure nobody uses UTF-8 or even entity names as
its canonical processing encoding... That would be a nonsense. But who
knows, masochism exists, I know (:

Alain LaBonté       (version : 8 bits ---  (-:                         )
Alain LaBont/e'/    (version : 7 bits ---  )<:= !@#$%?&*()_+-=^~',."!!!)
Received on Wednesday, 5 February 1997 11:29:14 GMT

This archive was generated by hypermail 2.2.0+W3C-0.50 : Tuesday, 2 June 2009 19:16:46 GMT