Re: iso-8859-1-Windows-3.1-Latin-1

[ Taking Liam out of CC since I know he's on w-v. ]

On 08.05.01 at 21:41, Liam Quinn <liam@htmlhelp.com> wrote:

>On Wed, 9 May 2001, Terje Bless wrote:
>
>>Hmmm. Maybe Björn or Sean can clarify a bit, but I don't really see any
>>big problems provided the charset used is duly registered with IANA and
>>marked as suitable for use as a MIME encoding.
>
>The user agent may not support the encoding, as is commonly the case with
>windows-1252 on platforms other than Windows and Mac.

Yes, this is a good argument for authors to limit their use of esoteric
charsets. However, the spec does punt available charsets to IANA and IANA
refers to RFC 2978 <URL:http://www.ietf.org/rfc/rfc2978.txt> and
<URL:http://www.iana.org/assignments/character-sets>. A User Agent that
does not expect this and deals with it in a reasonable manner cannot be
said to be fully compliant with the spec. IMO, of course.

User Agents have considerable difficulty with UNICODE!

Having just checked the WCAG, it does not deal specifically with character
encoding issues. The general sense of it is the same as the general case:
use the minimum level of technology that will get the job done. For charset
issues, this would be ISO-8859-* where possible and UTF-8 otherwise. In
particular, windows-1252 is actually an acceptable compromise in this case
because it is more accessible then the equivalent UNICODE; cf. WCAG
Guideline #10 <URL:http://www.w3.org/TR/WCAG10/#gl-interim-accessibility>:
"Use interim solutions".


It's also worth noting that the WCAG has completely different priorities
then common wisdom. It explicitly takes into account browser bugs because
the bugs exist and must be dealt with in the real world. This as opposed to
our little Ivory Tower here where we disregard _all_ implementations in
favour of a theoretical model of how this /should/ be implemented.

Received on Tuesday, 8 May 2001 22:14:48 UTC