>It seems I asked practically the same question about font display back in July 
>this year, to which Smokey Ardisson replied:

>>Without seeing the website and the fonts it's hard to say for certain, but 
>>there are a number of bugs where Gecko doesn't recognize certain fonts, or 
>>glyphs within fonts, on the Mac.  You could be hitting one of those bugs.

>Is that really all the information I can get?

Since you provided a website here, I took a look at it :-)

The site you provided <http://www.learntibetan.net/helpfont.htm> is written in 
such a way that requires using an old, non-standard, pre-Unicode font*, which 
is pretty dodgy these days.  The web and its browsers are moving rapidly to a 
Unicode-based model, away from legacy standards and certainly away from 
non-standards random fonts like those employed by this website.

That said, the issue here doesn't seem to be a hole in Gecko's support for 
certain scripts as I mentioned in July.  It seems Gecko is simply failing to 
recognize the "TibetanMachineWeb" font and is only recognizing the 
"TibetanMachineWeb1" font (the downloaded font file consists of 10 separate 
fonts inside).  It may be bug 246527 
<https://bugzilla.mozilla.org/show_bug.cgi?id=246527> (which does seem to 
strike non-Roman languages disproportionately) or it might be a bug in the way 
the fonts were made.

>That if I want to look at any of the websites that might, just might, be using 
>those Tibetan fonts (or presumably any of a great variety of other language 
>fonts), then I had better be using another browser? That's a real 
>disappointment! Is there any hope on the horizen?

There are two pieces of good news.  First, it seems like Gecko does a decent 
job handling real, standards-compliant Unicode Tibetan, at least as far as I 
can tell from Alan Wood's test page 
<http://www.alanwood.net/unicode/tibetan.html>.  So sites that are done using 
Unicode Tibetan should work.  With respect to other languages and scripts, 
generally the holes are pretty small, like not being able to set certain fonts 
as default in the preferences or some characters showing up in another font 
(but as the right character).  The worst case, as far as I'm aware, is the 
support for South Asian languages; Gecko on Mac OS X currently can't do the 
glyph reordering that's required (it needs ATSUI).  And probably Mongolian, 
since it requires contextual shaping like Arabic AND is written vertically--but 
that's a challenge for everyone yet!

Second, and more generally, the next big step for Gecko on Mac OS X is to a 
new, modern architecture for graphics and text-rendering.  No more QuickDraw, 
no more whatever we're using for text now (maybe QuickDraw for that, too); 
hello Quartz and ATSUI (the same text drawing routines used in Safari and most 
of the OS, with really top-notch support even for less-common scripts.  It's 
just slow unless you feed it text in a certain way, which Gecko's not set up to 
do yet.)  After the latest series of releases come out, Josh will be locked in 
a closet and made to develop lots of this new code.  Or something like that; 
obviously Mike and Simon and Josh are the ones to ask here :-)

Sorry, you didn't ask for a novel (or a lecture), but I'm afraid my reply has 
become one, if not both :-(  I hope I was at least somewhat informative rather 
than confusing and answered your questions.

Smokey

* In Unicode, every character in each of the world's writing systems has its 
own "codepoint", so each can be written and shared unambiguously across the 
Internet (and all on the same webpage, if one wanted).  Prior to that, people 
often made fonts where the replaced the Roman characters with the ones from 
another writing system, so that the codepoint for, say, a, was used for the 
Hebrew letter aleph. In pages which specified that particular font, text 
appeared as the author intended, but the actual text itself was gibberish in 
Roman letters.  Not good for unambiguous interoperability and data exchange, 
because proper understanding depended on appearance, and appearance depended on 
the author and all recipeints having the same non-standard font.  And some 
other, more complex issues :-)  This was particularly common for "less common" 
languages and writing systems, but Mac OS versions before 8.5 or 9 did it for 
Greek (with the Symbol font).
_______________________________________________
Camino mailing list
[email protected]
http://mozdev.org/mailman/listinfo/camino

Reply via email to