An architecture for the shaping of Indic texts

There has been virtually no software localization into any of the major lan guages of India. One important reason for this is the fact that enabling In dic scripts at the base software level involves sophisticated computational process that are far from the traditional font level substitutions that su ffice for a number of other world languages. Indian languages are basically phonetic in nature and fortunately, their writing systems are amongst the most logical and are thus highly amenable to being embedded in software. Th is paper is primarily concerned with the shaping process needed at the kern el level of the operating system so that software systems can include suppo rt for enabling Indic scripts. The shaping architecture and computational p rocess described are based on over two decades of work in trying to build b asic support for Indian languages in computer systems. We first present the basic phonetic nature of Indian scripts and the unique characteristic natu re of writing in Indian languages. Next the computational process of shapin g and a general architecture for its implementation are described. Finally the specific implementation for Unicode encoded texts displayed using TrueT ype Open fonts is briefly presented. (C) 1999 Elsevier Science Ltd. All rig hts reserved.