Re: HTML forms and UTF-8

From: Erik van der Poel (erik@netscape.com)
Date: Mon Nov 15 1999 - 12:49:37 EST


[I am resending this, since it didn't appear to make it last time...]

Tim Greenwood wrote:
>
> [default font problem...] This flaw may be corrected now, but I
> still do not like the way that Communicator 4.6 handles UTF-8. The Edit
> Preferences form has you set fonts for 'Encodings' Most 'Encodings' are
> given appropriately in terms that users will understand - Western, Central
> European, ... - but it includes Unicode.

The Communicator 4.X series is now being handled by a small team that
probably won't make any big changes in the Unicode and font areas.

Mozilla 5.0, on the other hand, is currently planning to use "language
groups" in the font preferences dialog. These language groups will be
Western, Central European, Japanese, Simplified Chinese and so on, again
mostly aligned with the traditional character encodings. We will be
looking for the LANG attribute in HTML and the Content-Language header
in HTTP, and using that language info to decide which font pref to use.
(This solves the Han unification problem, by the way.)

Most documents lack the language info, so we plan to use the charset as
a fallback. E.g. Shift_JIS probably means that the LANG is ja.

We will have 2 exceptions in the language groups: Unicode and
User-Defined. These are actually charsets, not language groups. The
current thinking is that documents labelled with a Unicode charset are
more likely to use text that cuts across the boundaries of the
traditional charsets (e.g. mixtures of Latin-1 and JIS characters),
therefore requiring a larger font so that font switching (aka font
linking) does not occur too much. (Font switching will still take place
if characters requiring glyphs outside that font occur.)

The User-Defined font pref will re-appear to cater to those users that
are used to using that in the older browsers. The User-Defined fonts
will be used when the user has set the charset menu to User-Defined (as
usual).

Comments are welcome!

Erik



This archive was generated by hypermail 2.1.2 : Tue Jul 10 2001 - 17:20:56 EDT