Recently there has been a spate of projects which have released classical Sanskrit and Tibetan E-texts. There appears, however, to have been little or no collaboration between each project and this has resulted in a growing number of transliteration schemes. In this paper we describe these schemes and their differences and propose a method by which they can be distinguished automatically.
  We begin by separating the various transliteration schemes and then extract from them various features through which identification becomes possible. With Sanskrit, it turns out to be extremely difficult to identify all the schemes by this method alone. In consequence we rely on sandhi, a sound phenomenon in Sanskrit to increase the probability of identification. This foundation can form the basis for software identification of the various styles of E-text.