|
Computer fonts for
the Khmer language have existed now for almost 15 years. They were
not developed using Unicode, as Khmer Unicode did not exist at the
time. These fonts are now being widely used for word processing as
well as for newspaper and magazine typesetting. Keyboards for these
fonts exist and are used.
The
Unicode coding of Khmer has
become considered finalized by the Cambodian government (except for
possible bugs) in the version 4.0 of Unicode (it was already
included in versions 3 and 3.1). Unicode not only
establishes which code is assigned to each letter of the Khmer
alphabet, but also specifies in which order they are typed, which
changes from the prior systems. In handwriting and with non-standard
fonts, Khmer was typed from left to right, in the same order that
characters are printed. With Unicode, Khmer is typed in the order
that letters are pronounced (different), which requires an
"intelligent behavior" from the system or word processor.
New font formats
(OpenType, ATT) permit the coding of this "intelligent behavior"
inside the fonts, allowing them to be used in standard systems,
without having to develop much specific code for them.
A standard for Khmer
OpenType (developed
by Microsoft) and some OpenType fonts for Khmer already exist,
codifying all the necessary information. Most of these fonts are
proprietary. The KhmerOS initiative has already produced and
placed in the public domain a full Unicode OpenType font family, in
order to allow for the development of OpenSource software that
interprets Khmer fonts. The project is presently working on
improving this font, hinting it, so that it can be used as a system
font, visible in very small sizes. Please see the
status and
download pages for news on
fonts produced by the KhmerOS initiative or other sources.
A full set of fonts has to be bought
or developed and put in the public domain, it has to include
basic (very clear) fonts to be used in menus and other parts of the
user interface, as well as other fonts in which the artistic side is
more important, to be used in documents and design. The set of fonts
has to be an improvement on the non-standard sets used now to do
Khmer text processing in English-language based computers, so that
fonts are not a barrier in the deployment and acceptance of the
system (reducing the cost of change for users who already type Khmer
in other systems).
Fonts are a particularly important part
of this project, as they can add value that other systems may not
have. The richness in fonts of the Macintosh system turned Apple in
the 1990s into an industry reference for publishing, owning a lot of
its merit is to enormous variety of fonts.
The sets of fonts developed will have to
cover very different needs. Some of these fonts will be bought, some
will be developed and some will be adapted to Unicode and OpenType,
re-using efforts made earlier by font developers, computer
manufacturers and scholars:
-
Fonts for computer screens. Very
clear fonts that are easy to see in a computer screen, even at a
very small size. This fonts will be used to translate system and
application messages. A full family of fonts is necessary
(normal, bold, cursive, cursive-bold). The fonts need to be well
“hinted” for their use in small text. We actively working on
this font at the moment.
-
Fonts for word processing. These
fonts must include traditional Khmer fonts, old style fonts
(ancient scripts), fonts for signs (Mul), square fonts (Chrun),
handwriting fonts and newer (fancier) typefaces.
-
Fonts for designers. Present fonts
are not well fitted for design, aesthetically as well as
technically. New designer fonts must be developed, assuring that
they can be used and “bended” on the ways designers have to.
Styles from western and far-eastern countries should be
considered when designing these fonts.
-
The full list of fonts should
include:
-
Representation of historical texts
in their original style.
o
Brahmi script (1st
to 5th centuries) (1 model).
o
Khmer ancient pre-Angkor
script (6th to 9th century) (1 model).
o
Khmer ancient Angkor
script (10th to 13th century) (3 models).
o
Post-Angkor Khmer Script
(14th to 19th century) (many models).
o
A full family: normal,
bold, cursive, cursive-bold.
o
Fonts for Signs (Mul).
o
Square fonts. (Chrun).
o
Handwritten script.
o
Fancy typefaces based in
Khmer, European and Chinese/Japanese calligraphy.
The fonts for this project must conform
to industry standards, and use advance typographical formats that
are well prepared for Indic scripts such as Khmer. Both Microsoft
and Apple have such formats (OpenType for Microsoft and ATT for
Apple). The fonts must also use the Unicode standard.
A number of these fonts (15) are already available in OpenType
format, but they are copyrighted. Part of these project would be to
try to buy them into the public domain.
In total the initiative should buy or produce at least 50 fonts or
variants for the public domain.
Om Mony, Cambodias pioneer
typographer, also makes Unicode fonts available through his
www.camboday.com website.
As of August 2004, we consider that
enough Unicode font exist, and, unless there are funds available, it
is no longer the business of the project to try to bring to the
public domain more fonts, unles specific funding becomes available.
Fonts produced by Danh Hong, Om Mony, the Khek brother and other
sources are considered sufficient. Nevertheless, we do want to be
careful about fonts complying to the standard. Fonts that include
specific ligatures instead of characters (such as using the word AUI
instead of the independent vowel QUUV2) have been detected. The use
of these fonts produces incorrect Khmer Unicode texts, and should be
strongly discouraged.
Translation of text in legacy
fonts to Unicode
It is quite reasonable to think that
-with the use of Unicode- all old fonts will be phased out.
In order to reduce
the cost of changing from an old system to this one, it is important
that files and documents written using old Khmer fonts (pre-Unicode)
be translated easily to Khmer fonts. An application to translate old
documents need to be developed.
There already
exists a utility, developed by Jean Yves Fusil and maintained by
OpenForum to translate texts written in one type of old (legacy or
private encoding) Khmer computer font to another old type of Khmer
font. This program –able to translate among 23 codifications of
Khmer fonts- does not handle Unicode at present.
Lin Chear developed a first version
of a ABC to Unicode converter, taking a plain text file and creating
a scaped Unicode file. The we decided to get into it and created a
more complete converter that transfers a plain text ABC file (taking
into account Zero Space) into a Unicode utf-8 file. A second version
of the program that starts directly from a Limon font has also been
made available. Both programs are available from our
download page.
Their characteristics are:
-
Differentiating
between SRA U and the lower forms of TRAISAP and MUSIKATOAN.
-
Recognizing all of
the cases in which PO + SRA A are used instead of the letter NYO.
-
Getting most of the
cases of Coeng TA and Coeng DA correctly.
-
Reordering Register
Shifters and diacritics when they are located in the wrong
place.
-
Reordering Vowels,
Signs
and Coeng consonants when they are in the wrong order.
As for document
file format (passing from MS Windows to Open Source environment), no
translation program is necessary. Even if the Microsoft format for
documents (UCS-2) is different from the format used by Linux
(UFT-8), the applications do this translation automatically when
they open a document created with Microsoft programs (such a MS Word
or MS Excel).
Check our status and
download pages for updates. |
