Using Unicode fonts with Mac OSX
[February 2, 2010]

Standard Unicode fonts contain all the diacritical marks that are needed for reproducing most Indian languages in transliteration (at least the languages with which I work most often: Sanskrit, Kannada, Tamil, Prakrit, Pali). An additional advantage of using Unicode rather than the older coding schemes is that the texts can be read and processed further on any computer with an operating system that supports Unicode: not only Mac OSX, but also Linux, and the later versions of MS Windows.

For the successful Indological use of Unicode with Mac OSX, one needs
  1. a Unicode keyboard driver
  2. at least one Unicode font
  3. application software that supports Unicode

Problematic issues

Unicode fonts with all the necessary diacritical signs which are needed for Indological writing come in various kinds, for instance, TrueType, OpenType, AAT (Apple Advanced Typography), and in the Mac-specific dfont format. In general, Mac OSX supports all these kinds of fonts.

However, often there are difficulties when we Indologists wish to use fonts for Indian scripts. OSX comes with fonts and keyboard drivers for (Deva)Nagari, Tamil, Gujarati and Gurmukhi. These fonts have been made specifically for the font rendering engine that is a part of the OSX operating system. This engine (called ATSUI - Apple Type Services for Unicode Imaging) determines how combinations of signs in complex scripts are produced correctly (or not). For instance, the combination 'consonant + short i' in north Indian scripts should be rendered with the sign for the short 'i' coming first, being followed by the consonant sign. If one attempts to use an OpenType or TrueType Nagari font that was designed for use with Windows™, the rendering will most likely be incorrect (unless one uses XeTeX - see below).


Keyboard drivers

A good keyboard driver for typing the Indologically necessary diacritics, which is included with OSX, is, miraculously, "Inuktitut-Nunavut". "U.S. Extended" does practically the same, but apparently did not allow one to type a long vocalic r under OSX 10.3; however, it does under OSX version 10.5, with the handy and totally unintuitive combination ALT-a + r, just as the Inuktitut layout (thanks to Marco Franceschini, Bologna, for pointing this out). Dr. John Smith, Cambridge, offers a keyboard driver via his FTP site (here, where good fonts can also be found - see below), which is very convenient to use for Sanskrit; it also has the necessary signs for Dravidian languages (to be composed, using 'dead keys': alt-= for a macron over Dravidian long e and o, and alt-+ for the underbar with l, r and n).

If one wishes to create one's own keyboard driver, SIL International (Summer Institute of Linguistics) offers a free piece of software for doing so named Ukelele.

Keyboard drivers for Kannada, as well as an improved version of the Kedage font (the only one that functions well under all circumstances with OSX), can be found here.


Fonts

A standard installation of Mac OSX contains a few such fonts: Times, Palatino, and Helvetica. Other Unicode-enabled fonts, like Lucida Grande and Monaco, lack certain styles.

Elegant professional fonts that contain the necessary diacritical marks, available in versions optimized for OSX and for Windows, are offered by SIL International, such as Gentium Basic, which also comes in a slightly heavier version called Gentium Book Basic. The older, original Gentium (still available) did not offer bold face. A newer and very nice font is Charis SIL.

The Unicode fonts offered by Dr. John Smith of Cambridge at his FTP site: New Century Schoolbook, Palatino, Times, Courier and Helvetica, are a reworking in OTF (OpenType) form of his older TTF (TrueType) fonts, that were based on original fonts made by URW++ Design and Development Incorporated in Hamburg, Germany. However, some users may find that they are not so satisfactory on older systems (e.g., I had some problems on my PowerMac G5 when I still used the old OSX 10.3.9 system) as the earlier TrueType fonts: no underdotted, overdotted and underscored letters are produced. Since the old fonts are no longer available on Dr. Smith's FTP site, I am making them available here: New Century Schoolbook, Palatino, Times, Courier and Helvetica. With the first three of these five fonts, printing bold and italic text is also possible.

Much information about available Unicode fonts (also for Indian scripts), and about how to install them, is given by Alan Woods on his "Unicode fonts for Macintosh OS X computers" page.

A test page in RTF format with a speciment text in transliteration in the above-mentioned fonts can be downloaded here, and the PDF version of the same here.

As has been mentioned above, there are special technical problems concerning the use of fonts for Indian scripts. Besides the Indian scripts that are supplied with OSX (for Nagari, Tamil, Gujarati and Gurmukhi), I know of one font that is quite good for Kannada, namely Kedage.


Application software that supports Unicode

Not all OSX applications support Unicode equally well, some of them poorly, and some not at all.

Word processors

I have difficulties using Microsoft® Word 2004 version 11.0: the program does not recognize some fonts, and at times texts that were written using other word processors, in which all required diacritical marks appeared correct, are rendered with odd distortions and 'empty boxes'. Far more satisfactory is the freeware NeoOffice, which is not just good and can open all the usual M$ Office formats, but is also available free of cost, based on the open source OpenOffice, of which there is a version for OSX that runs under the X Windows interface and lacks some of the user comfort that NeoOffice has. (As per July 2006, a native version of OpenOffice is in the making.) The very reasonably priced Nisus Writer Express (and its enlarged version, Nisus Writer Pro) at present seems to give the best allround Unicode support. Also Mellel was developed with the handling of non-European scripts in mind, but at present it does not yet support fonts for Indian languages.

If one also wishes to use Indian scripts in one's texts, NeoOffice is good for Kannada but has (as per version 1.2.2) an odd problem with the Devanagari font: when an anusvara is placed, some of the secondary vowel signs above consonants (e and ai) are moved leftward. - Nisus Writer (Pro or Express) seems to performs best.

As a robust and highly adaptable alternative to these word processors, the TeX typesetting system must be mentioned here (and especially the extension for OSX [and for Linux; there is also an experimental version for Windows™] called XeTeX, which makes it possible to use any installed OSX font in TeX). TeX is no longer a thing for computer hobbyists, technicians, nerds and geeks, but has actually become easy to use, thanks to the various editors and graphical frontends that are in existence now. Entirely free of cost, one can have an excellent, highly stable, powerful text processing system that delivers beautiful printed results and is platform independent (unless one uses platform-specific features). For examples of results in print, one can see the publications of the Institute of Indology and Tamil Studies in the University of Cologne, and the writings of Dr. Dominik Wujastyk, the Indologist who has contributed some items of interest to the ever-expanding TeX system. One would wish that more scholars and publishers use the versatile TeX system.


E-mail and Unicode

E-mail clients for OSX that support Unicode include Apple Mail (which is included with OSX), the mail agent of the freely available, beautiful, innovative multi-platform Opera web browser, and the superb free, multi-platform Thunderbird.

However, Thunderbird has difficulties correctly rendering Indian scripts. (I have tested Kannada - the classical-looking Kedage font by Mr Nicholas Shanks, based on a font from Bangalore, which is the only one to date that functions well under OSX - and Apple's Devanagari font.) Not entirely surprisingly, among these e-mail programs Apple Mail and Opera (since version 9.50, which has brought a dramatic improvement) perform best with these Indian scripts. (As a workabout solution, messages that are received in Kannada or Nagari with Thunderbird and that look ugly, can be copied by means of cut-and-paste into Apple's little editor TextEdit, included with OSX, where they appear correctly.)


Web browsers

Support for Indian scripts has significantly improved with the standard Apple browser Safari under OSX 10.4 (Tiger) in comparison with 10.3 (Panther). Taking the Wikipedia in Kannada, Tamil, Sanskrit and Hindi as a test case, all the pages are excellently readable. As of version 9.50, also Opera renders them well (also under OSX 10.3.9). Other browsers, like Firefox, work fine with Latin script with diacritical marks in Unicode, but there still are some difficulties rendering Indian scripts.


XeTeX

Jonathan Kew at SIL has produced a version of TeX / LaTeX that, astonishingly, directly uses any regular font installed on a Mac OSX computer (version 10.3 and higher), called XeTeX. This too is freely downloadable and usable, and it is included in some distributions of the TeX system for Mac (such as the excellent MacTeX, free software which can be found on CTAN (the Comprehensive TeX Archive Network) and its mirror sites, as well as on the TeXLive distribution on CD, which can be ordered from various TeX distributors and also includes the software for Windows systems). - XeTeX supports Unicode, and using a freely available (La)TeX editor such as TeXshop or iTeXMac (included in MacTeX), one can input text with diacritical marks straight into a TeX source file, from which a PDF file is produced, that can be printed and viewed on all major hardware platforms: Mac, Linux, Windows, and more. Recently a version of XeTeX has been brought out for the Linux operating system too, and there is also a port for Windows (link to a Japanese site via the SIL XeTeX page). I have not yet tested the Linux version; the original for Mac is delightfully simple to use, and also supports the use of the fonts for Cyrillic and Devanagari and other non-European scripts that come with OSX. - Furthermore, XeTeX makes it possible for OSX users to use a much larger variety of fonts for Indian scripts, because it has its own font rendering engine that renders TrueType and OpenType fonts correctly.

The MacTeX TeX distribution comes with a great deal of documentation, in different languages. It can be found here, and further documentation can be found here. Another excellent distribution for Mac, which is more easily customizable for those who have specific demands and do not need everything that comes with MacTeX, is gwTeX, which is easily installed by means of the i-Installer. A very informative web page about using TeX on OSX is this one, by Gerben Wierda, the maker of i-Installer.


TeX editors / frontends and macro packages (LaTeX, ConTeXt)

TeX and LaTeX (i.e., TeX with the standard macro package by L. Lamport, which makes numerous tasks in the TeX typesetting system easier and has for many become the usual way to use TeX) are used most easily and efficiently on the Mac by means of a graphical frontend: an editor that includes further macro functions. Some of these editors are downloadable from the internet and may be used free of cost. My personal favourite is TeXShop, available here and elsewhere. (TeXShop can also be used extremely easily with ConTeXt, the modern preprocessor-and-macro package that is increasingly seen as an interesting alternative to LaTeX, and ConTeXt and XeTeX can also be combined easily.) Another good editor is iTeXMac, available here and elsewhere. (iTeXMac can also be used with XeTeX, and SIL offers an installation script here. This is a bit more cumbersome than using TeXShop, but it works. And it works easily with (Xe)ConTeXt as well.)

ConTeXt is very consistently structured, is easily learnt, and in my opinion allows the user to alter the finer details of TeX typesetting more easily than LaTeX does. This may be largely a matter of personal requirements and taste, but in any case it is worth looking at. Just like XeTeX, ConTeXt is included in MacTeX, and detailed documentation can be found at the ConTeXt website. There is also a wiki with many links to further relevant pages. A web page in German explains the main differences between ConTeXt and LaTeX with numerous concrete (very simple) examples (source files and PDFs).



Home (German)

Home (English)