Statement of Purpose:

The primary purpose is to present a unified phonemic code based technique that can be used to support Indian languages in a variety of applications. The complexity of processing Indian languages is first presented. This is followed by a detailed exposition of this phoneme basis, its ancient historical background, applicability to Unicode and ISCII, a unified technique for Indian language processing tasks, with specific examples from a shaping/rendering engine using OpenType fonts. Finally experience from varied applications is briefly discussed.

Brief Description:

The multitude of Indian languages and dialects are written using 9 scripts. While these scripts have been allotted distinct code pages in the Unicode scheme, applications supporting Indian languages are yet to be found on a number of standard platforms. One primary reason could be the fact that rendering, and processing in general, of Indian languages is complex and mandates distinctly different techniques. Orthograpy follows a phonetically driven basis of compositing "phonetic units" to form complex glyphs. While the character set is compact, authentic rendering implies a generative mechanism that can produce glyphs corresponding to all possible character sequences. Complex as it may seem, clear rules can be defined based on a canonical treatise by Panini, the ancient grammarian. These rules establish a perfect correspondence between phonemes constituting a syllable and its graphical form. And such rules can be defined for each of the Indic scripts. Decomposing text using this phonemic basis, followed by phoneme based computations provides a single unified technique for rendering Indic scripts. In fact, it is well suited even for other processing tasks such as sorting, searching, speech synthesis, speech recognition, transliteration, etc.

Conclusions:

The software complexity of supporting Indian languages in different applications can be controlled by the use of a unified technique based on phonemic codes obtained from a well defined transformation of Unicode or ISCII encoding. This is amply illustrated by actual implementation experiences.