Computers & Writing Systems
NRSI Update #1 – July 1996
Welcome to the first issue of the NRSI News Update!
In this issue:
What is the NRSI?
by Margaret Swauger, Manager NRSI
The Non-Roman Script Initiative (NRSI) is a project of SIL’s International Publishing Services focusing resources on issues pertaining to using, displaying and researching non-roman (NR) scripts. There is an NRSI-Central team in Dallas, but NRSI team members are working in field locations. Among other activities, we develop NR fonts and scripting systems, keep up to date on current and emerging technologies that benefit the NR script user, and coordinate NR information.
New KeyMan Beta 3.2
by Wes Cleveland
A new version of KeyMan (version 3.2) is currently in Beta Testing. This version contains several bug fixes and enhancements including:
$keyman: version 3.2 $silkey: version 1.1
would mean this .kmn file is compatible with KeyMan 3.2 and SILKey 1.1.
Although we are nearing the end of the Beta Test program, new Beta Testers would be welcomed (if you’re serious about testing and don’t mind an occasional disruption).
by Dennis Drescher
One of my responsibilities is to look into new publishing applications that would be suited for non-roman minority script languages. I’ve had an opportunity to work with three applications that work with GX fonts: “UniQorn”, “Ready, Set, Go!”, and “TeXgX”.
From these initial investigation, it’s clear that UniQorn and Ready, Set, Go! are not suited for long document publications. However, they are well suited for work we do with literacy and other shorter documents. In my opinion I think they could handle most scripting systems but time will tell. I’ll be working with these two applications more during the summer and should know better what their limitations are by fall.
As for our long-document publishing, TeXgX is the only viable tool available in the non-roman environment. It’s not what you would class a user-friendly application but it works better than anything else I’ve seen. Jonathan Kew has done a fantastic job of porting TeX over to the GX environment.
by Victor Gaultney
Apple’s WorldScript (WS) technology, introduced four years ago, has been used successfully to provide non-roman script solutions inside and outside SIL. Preparing WS “Language Kits”, however, has always been a difficult programming task, discouraging their development. Only a few applications were “WorldScript- savvy” and supported custom kits. Apple was also reluctant to make the WorldScript system extension available outside their own kits.
WorldScript provides smart font rendering capabilities for non-English languages and scripts on the MacOS. Although this is limited to only a subset of those available with QuickDraw GX, WS also provides word demarcation, collating, tokenizing and date/time/calendar features — capabilities that are beyond the scope of GX.
With the WorldScript extension and a Language Kit (a package of fonts, utilities and a WS module) for a particular script, a user can work in that script on the standard US version of the MacOS. Apple has released Language Kits for Chinese, Japanese, Arabic, Hebrew and Cyrillic, with Korean, Indic and others in development.
Recent developments have changed the status of WorldScript and have made Language Kit development more attractive for SIL:
by Peter Martin
We are planning to set up a list server to provide a forum for discussion among those in SIL and cooperating agencies involved in non-roman script development issues.
Participants will be able to send a message to the list server, which automatically distributes it to all other list subscribers.
Watch this space for further news.
Articles submitted to NOC/NOAM:
TrueType Open - Update #2
by Bob Hallissy
Research into Microsoft’s TrueType Open technology is proceeding on three fronts: understanding the facilities available in Middle East Windows 95, dialog with Microsoft about the future of TTO such as the new APIs, and design tools to facilitate the building of TTO fonts.
TrueType Open (TTO) is a Microsoft designed extension of the TrueType technology that provides facilities such as contextual substitution and glyph positioning. We use the term “smart font” to describe technologies such as TTO and Apple’s QuickDraw GX fonts because there is the ability to encode information into the font itself that modifies the way each character is displayed based on its context. Such capabilities allow us to separate decisions about how textual information should be stored in the computer from decisions about how the text should appear on the screen.
For example the Ethiopic writing system is a syllabary that includes separate symbols for almost all combinations of consonants and vowels. Because there are around 500 such symbols, most existing Ethiopic solutions rely on using several fonts and having the application change fonts to get the needed characters. Smart font technologies allow us to store Ethiopic text in the computer file as simple sequences of roman (or phonetic) characters _ the smart fonts know how to combine consonants and vowels into the appropriate Ethiopic symbols.
For further information about TTO, Microsoft has published TrueType Open specification and some rudimentary tools on their Internet site (no longer available).
You can also request the following NRSI documents by sending an email request to . Include the filename of the document in the body of your email message.
The NRSI approach to TTO
Though TrueType Open is still in development at Microsoft, it holds a promise of helping us deal with many non-roman writing systems on the Windows platform, and thus the NRSI is committed to research and development of this technology. NRSI efforts have proceeded along three fronts: research into Middle East Windows 95, discussions with Microsoft about future plans for TTO, and initial design work into a TTO font compiler.
Middle East Windows 95
Only two of the promised “localized” versions of Windows 95 are still in beta test: the Arabic and Hebrew versions. These two are built on the core product that is known as Middle East Windows 95. This core product is of interest because it implements a subset of TTO and because it handles mixed direction script. Because certain writing system behaviors (such as contextual forms) are encoded into the fonts instead of being completely hardcoded into the operating system, we may be able to create fonts for minority languages.
The unknowns relate to how much flexibility exists within the subset of TTO as implemented. For example, can we change a character’s behavior from being a “joiner” character to a “non-joiner?” Or, can we change a character from being a diacritic mark to a base character? Once we know the answers to these kinds of questions, we can determine whether a specific non-roman script project can or cannot be implemented using the Middle East Windows 95 platform.
Just recently we have successfully implemented a font that triggers some contextual substitution behavior in a beta release of Arabic Windows 95. We used SIL’s TypeCaster program to build a font, and then used Microsoft’s TTO “assembler” to imbed some rules into the font. This process is not for the faint of heart as the assembler is extremely rudimentary and detailed knowledge of TTO font structures is required. But with this success we can now begin to see just what flexibility exists in the Middle East platform.
The Future of TTO
In order for TTO to be useful beyond the limited subset of TTO that is incorporated into Middle East Windows 95, application developers will need to include full support for TTO in their applications. It is our belief that it will have to become a lot easier to support TTO than it is today before developers will choose to incorporate the technology into their applications.
Microsoft is working on a set of APIs that is supposed to enable solution developers to capitalize on TTO, but we have not yet seen any concrete information about these APIs in order to make an evaluation. We had originally expected to see the definitions of these APIs this past spring, but we have not heard anything yet.
TTO Font compiler
In addition to having applications that support TTO technology, we must be able to create TTO fonts. The tools provided by Microsoft so far are very low-level and difficult to use. They require detailed knowledge of TrueType Open font structures. We are presently working on the design of a higher-level tool that will make it reasonable for a script engineer to implement TTO-based solutions.
This work is in its infancy and will most likely be implemented in stages. The first stage will be a text-based program where the TTO data is coded using any standard text editor. A subsequent program will provide a more friendly front end to the process.
Circulation & Distribution Information
The purpose of this periodic e-mailing is to keep you in the picture about current NRSI research, development and application activities.