Unicode fonts in Browsers (was Re: What will happen in the future?)
Patrick T. Rourke
ptrourke at mediaone.net
Tue Mar 6 08:14:34 EST 2001
> We're in better shape on the Greek side since the SIL Galatia font
combines
> the vowel and accent/breathing into ONE character (when used with
Keyman --
> of course, you could always do it the hard way of using the <alt> key plus
> the numeric keypad). Unfortunately the browsers don't support polytonic
Greek
> in Unicode.
I'd be interested to know where you obtained this misinformation.
The following TrueType (or in one case OpenType) fonts known to me support
polytonic Greek with precomposed characters (i.e., both the basic Greek and
Extended Greek ranges of Unicode), and work in Internet Explorer 4.0 and
higher and Netscape 4.0 and higher and Mozilla M14 and higher in Microsoft
Windows 95, 98, 98SE, ME, and 2000 at least (and probably in NT4), with
Mozilla M17 and higher in current version versions of Gnome, and with
Konqueror in KDE 1.96(?) and higher.
Aisa Unicode (comes with Multi Key)
Arial Unicode MS (comes with Office 2000)
Athena (freeware)
Cardo (freeware, I think)
Code2000 (shareware)
Georgia Greek Unicode (freeware)
Palatino Linotype (comes with Windows 2000; OpenType)
Titus Cyberbit Basic (licensed for non-commercial users)
Vusillus Old Face Italic (comes with Antioch)
[Listed in alphabetical order. The best looking freeware fonts in my opinion
are Aisa, Cardo, Georgia, and Vusillus (in no order). All except Arial are
serif fonts.]
The following support polytonic Greek via combining diacriticals in the
above named browsers, except possibly in Linux. Of course the software to
properly place the Greek accents doesn't yet exist for either the Windows or
Linux operating systems. However, I have not yet managed to get the other
fonts to work acceptably in Linux.
Arial Unicode MS
Code2000
Lucida Sans Unicode (many versions)
Titus Cyberbit Basic
Vusillus Old Face Italic
[Lucida is a sans serif font] Several of the fonts listed in the first list
are being worked on to provide combining diacriticals support. Note that at
least one font (which isn't supposed to be used on Linux systems, and which
I removed from my Linux box as soon as I proved that it could be done) works
with combining diacriticals *even in Gnome and KDE* - though on my
underpowered Linux box, it takes ages to load a page using such a large font
(ok, literally, 8 minutes in Mozilla, 10 minutes in Konqueror).
Clearly U (a freeware BFD font, not a TrueType font) works in Linux, but I
find it hard to read because of the spacing.
Most of these fonts provide one case, but only one case, of the
Coptic-unique letters (but then again, most are not Coptic fonts: I'd be
very interested to know if anyone is aware of Unicode fonts with Coptic or
epichoric letterforms for the basic Greek range; obviously Coptic does not
require the extended Greek range)
If I state that a combination of browser, platform, and font works, it is
because I have managed to get it to work on one of my own machines and have
screenshots.
The Minion Pro font I'm sure works, but I haven't spend the money to buy the
font. And there are a couple of other fonts that I don't have which I know
work.
I imagine the Lucida Sans Unicode works in Charon for the Plan 9 operating
system; but I can't get Plan 9 to install on my system. Other OSs that
might be worth experimenting with are Mac OS9 with Netscape 4.5 or higher
(probably works, but I haven't been able to finagle enough time working on a
Mac to test it thoroughly), Mac OS X, and BSD.
My Unicode polytonic Greek pages are being thoroughly rewritten to discuss
NFC and NFD, to provide PERL conversion scripts between betacode and either
NFC and NFD (maybe JavaScript for ASP too), to provide some explicit
instruction on how to add Unicode polytonic Greek to web pages, provide a
number of sample texts (Euripides, hopefully a couple of inscriptions and
papyrus transcriptions; I suppose I could provide a chunk of John and a
small chunk of Coptic) and in general to shore the whole thing up;
unfortunately I've had a couple of recent setbacks in my work on the pages
(quick hint: if you try to install an experimental operating system on your
second hard drive, back up the first one!), and have been too busy to work
on them for the past several weeks. When the new version is done, I'll post
the URL, which will be different from the URL of the current (theoretically
non-public, and rather imperfect) draft.
As Peter Constable said, the preferred way to compose web pages (according
to the W3C character set recommendations) is with Unicode Normalization Form
C. NFC uses (roughly; see the Normalization Forms page at Unicode for
details; thanks to PC for first pointing out this page and the associated
white paper) the following characters: first, anything in the Latin-1 range
(i.e., mu from the micro symbol in the Latin-1 range); then, anything in the
Basic Greek range that isn't in the Latin-1 range (basically all the
unaccented characters and punctuation, and anything with only an acute
accent); then, anything in the Extended Greek range.
Many important classical literature projects are using combining
diacriticals for display: for instance, the Perseus Digital Library does so.
My machines are an Intel Pentium 133 MHz 64 MB 6GB box running (during
testing) RedHat Linux 6.1, 6.2, 7.0 and an Intel Pentium III 700 MHz 256 MB
30 GB /2 GB box running both Windows ME (on the large drive) and Windows
2000 (on the small drive). Linuces with versions of XFree86 lower than 4.0
are much harder to configure for Unicode TrueType fonts than those after
XFree86 4.0.
Patrick Rourke
ptrourke at methymna.com
(everywhere I've written "Unicode" above, it should be read as
"Unicode-compatible" or "Unicode-conformant" as they are not products of the
Unicode Consortium.)
More information about the Biblical-languages
mailing list