[WEB4LIB] Fonts

[Walt_Crawford at notes.rlg.org]

Andrew provides an answer that's certainly appropriate _for those pages
using a wide range of non-Roman scripts_.

Eureka does in fact use Unicode encoding to display a range of non-Roman
languages (Chinese, Japanese, Korean, Cyrillic, Arabic, Hebrew, Yiddish),
but those scripts don't appear to require the kind of typeface selection
Andrew discusses.

We've consistently found that, at least in IE (and apparently in Netscape
and Opera, but it's harder to be sure), the browser automatically switches
between the user-chosen typeface and the nearest Unicode-equipped
alternative as required by the text--even switching back and forth within a
single field (e.g., dates in an Arabic or Hebrew entry, where there's also
directionality switching). We were a little surprised by that, but it
helped our decision not to force typeface.

But of course, Vicnet seems to be dealing with a range of Unicode issues
that RLG hasn't had to deal with.

So maybe my question should be restricted to Web pages that only display
Roman1 and the rather large portion of Unicode that supports JACKPHY
languages... If you're using characters that require specific typefaces in
a manner that the Unicode placement won't define properly, it obviously
makes sense to force that typeface for that portion of the text.

-walt crawford-




                                                                                                                    
                                                                                                                    
                                                                                                                    






Walt_Crawford at notes.rlg.org wrote:
> Here's a question that has nothing to do with any particular
institution's
> home page (and I can stand flames...):
>
> For most material on most Web pages, why specify typeface (font) at all,
> either through CSS or directly?
>

There are practical reasons, esp when working with Unicode web pages.
Unicode encodes characters, sometimes these characters can be
represented by variant glyphs in different languages, making it
imporatnt to use different fonts for different languages.

Many users do not change the default fonts that their web browser uses.
If I was creating a web page in certain African languages, with the
character "Ŋ" (U+014A). I would not wnat to use the Times New Roman
font, since it uses the wrong glyph variant for many African languages.
If I was creating a page in the Yolngu-matha languages on the other hand
, I'd want to use a font that used the Sami variant (ie like Times New
Roman)

It becomes more problematic with Han Unification, where Chinese,
Japanese or Korean may use the same codepoint but use glyph variants.
For intsnace you're prepaing a web page in Japanese, and want to include
the Chinese names using the Chinese glyphs rather than the Japanese gylphs.

Syriac web pages are even more dependant on fonts. There are three
scripts (Estrangelo, Serto (West Syriac), and East Syriac), sharing the
same unicode block. If you need a particular script then it is necessary
to use an appropriate font.

Simarily, with the Arabic script there are two main style (Naskh and
Nastaliq). Most unicode fonts are Naskh, but as Nastaliq OpenType fonts
become available, it will possible to specific an appropraite font for
an appropraite language. For instance Naskh fonts for Arabic, or
Nastaliq fonts for Persian, Urdu, etc. Depending on typographic
preferences for different languages.

There are cases when for different languages, different fonts are
appropriate.


Andrew

--
Andrew Cunningham
Multilingual  Technical Officer
OPT, Vicnet
State Library of Victoria
Australia

andrewc at vicnet.net.au

Ph: +61-3-8664-7001
Fax: +61-3-9639-2175

http://home.vicnet.net.au/~andrewc/
http://www.openroad.net.au/