On Mon, 31 Mar 2003, Edward Cherlin wrote:

> On Monday 31 March 2003 04:31 pm, Jungshik Shin wrote:
> >Tomohiro KUBOTA wrote:
> > >I want such "alias" to be automated.  If I have one Korean
> > > font installed, it is obvious that renderer must use the
> > > font for all Korean texts. It is not a good idea that the
> > > renderer fail to display Korean when the user doesn't
> > > configure the "alias".
> >
> >     fontconfig always returns a font if there's a font on the
> > system with the character requested.
> > So, it's possible now.
>
> Doing it one character at a time is guaranteed to give hideous
> results. I have had the unfortunate experience of viewing a
> display in mixed CJK fonts, and I have had many similar

   Well, it depends on what kinds of fonts you have on your
system and the way you specify fonts you want to use. I'm well aware
of 'ransom note-like results when you mix up fonts of many *different*
styles and design principles in a single run of text.  This problem can
be minimized if you are careful in putting together fonts of similar
styles and design principles.

   Anyway, if someone finds it difficult to edit fonts.conf
file and doesn't want to install a minimal set of well-populated
fonts  (sans, serif, monospace, etc), but still wants
as many characters as possible to be rendered, randsom note
is what she deserve to get.


> unfortunate experiences of viewing APL code rendered in random
> math fonts. It is extremely important to a lot of people that
> they be able to specify a font *per language*, without regard to

  Well, *per-langauge* is not a cure-for-all although
on many occasions, it's sufficient.


> the definition of Unicode blocks or old-time code pages or
> ISO-8859-* or any other 8-bit font hack. But we want to do it

  We don't live in that world any more largely thanks to
fontconfig, Xft and Pango.  The age of X11 corefonts
and XLFD hack has gone for good.

> There is, of course, the question of defining the character
> repertoire and rendering rules for a language (which may differ
> substantially from the rules for another language written in the
> same script). To get started, it will suffice if I can say that
> the set of characters in one font that I designate defines the
> repertoire for my use of the language. When we have adequate
> support for more intelligent fonts, we can build in some of the
> rendering rules, also, but in the end language-specific document
> creation will be the job of applications well above the text

   In case of html, 'lang' does the job abd Mozilla supports
it pretty well. Unfortunaely, 'xml:lang' is not yet supported.


> editor level. At some point, explicit repertoire lists will be
> needed, I suppose. Or something else we haven't thought of yet.

   Care to take a look at http://fontconfig.org ?
It includes lang-dependent repertoire list for most, if not all,
of languages listed in ISO 639 (or is it ISO 30xx?)?


   Jungshik

--
Linux-UTF8:   i18n of Linux on all levels
Archive:      http://mail.nl.linux.org/linux-utf8/

Reply via email to