Michael, that is different than what I am talking about. There are two
different situations:
A. Program draws text with "MyComposite" font name. The OS goes to that
composition font, and uses the font list in the composite to break up the
text into ranges* and draws each range with the corresponding font.
B. Program (or library) determines which fonts installed on the system can
handle which of the characters given it, and breaks the text up into
ranges*, and draws each range with the right fonts.
* With proper BIDI ordering, of course!
There are pros & cons for each.
B depends on whatever heuristic the system has for determining which fonts
"go together". It also applears that B is not done automatically with the
basic drawing & measuring routines in Windows, but requires you to use the
Uniscribe library. I don't know if that library is always invoked when
drawing every piece of text that appears to the user: menu labels, button
labels, etc.
A has a specific list: there is no doubt as to which fonts you get for which
characters (unless some of the fonts are not installed). Moreover, every
time you use the composite font, even with the low-level string drawing and
measuring API, you get the same results.
Mark
—————
Δός μοι ποῦ στῶ, καὶ κινῶ τὴν γῆν — Ἀρχιμήδης
[http://www.macchiato.com]
----- Original Message -----
From: "Michael (michka) Kaplan" <michka@trigeminal.com>
To: <cfynn@druknet.net.bt>; "Mark Davis" <mark@macchiato.com>;
<unicode@unicode.org>
Sent: Thursday, October 18, 2001 9:29 PM
Subject: Re: Windows/Office XP question
> Actually, in Windows 2000 and even moreso in Windows XP, Uniscribe
achieves
> this, by properly doing both font substitution and handling of mixed
script
> usage (XP is a bit better than Win2000 as they added additional Indic
> scripts and also because they to a beit better of a job in sizing mixed
> script cases so you do not have the "normal sized Latin next to ultra-tiny
> Thai" problems that Win2000 could occasionally hit.
>
> This is not done by means of a font grouping, but more by having a
specific
> preferred font for every supported script any time the chosen font does
not
> support the script in question.
>
> MichKa
>
> Michael Kaplan
> Trigeminal Software, Inc.
> http://www.trigeminal.com/
>
>
> ----- Original Message -----
> From: "Christopher J Fynn" <cfynn@druknet.net.bt>
> To: "Mark Davis" <mark@macchiato.com>; <unicode@unicode.org>
> Sent: Thursday, October 18, 2001 8:58 PM
> Subject: RE: Windows/Office XP question
>
>
> >
> >
> > Mark Davis wrote:
> >
> > > One feature that
> > > some systems have is composite fonts, where the "font" is actually a
> table
> > > of subfonts in some order (perhaps with specific ranges assigned to
> each).
> > > That way, someone can have the advantage of specifying a single font
> name,
> > > and get a full repertoire, without requiring a monster font. Of
course,
> > > there may be little uniformity of style across scripts, or in mixtures
> of
> > > symbols, but at least you can get legible characters instead of boxes.
> > >
> > > Are there any plans to do something like that in Windows?
> > >
> > > Mark
> > > —————
> >
> > On some level at least this already seems to be implemented in Windows
> with system / GUI fonts. e.g. in Win 2K Unicode file-names etc are
displayed
> in the proper script in Windows Explorer if the system font for that
script
> is installed. There are seperate system / GUI fonts for each script,
rather
> than one huge font.
> >
> > A problem with implementing something which allows you to specify a
single
> font name and getttng a full repertoire is: Which font in script x matches
> font nnnn in script y? If I specify "Baskerville" for Latin text and that
> text contains a run of Arabic characters how does the system know which
> Arabic script text best matches Baskerville? Sure you could have a lookup
> table - but imagine getting users to maintain such a table with all the
> fonts some people accumulate these days. Font matching systems like Panose
> which might be used to automate this kind of thing seem to deal only with
> the characteristics of Latin and closely related scripts.
> >
> > - Chris Fynn
> >
> > --
> > Christopher J Fynn
> > DDC Dzongkha Computing Project
> > PO Box 122, Thimphu, Bhutan
> >
> > <cfynn@druknet.net.bt>
> > <cfynn@gmx.net>
> >
> >
> >
> >
>
>
This archive was generated by hypermail 2.1.2 : Fri Oct 19 2001 - 11:04:23 EDT