Apparently the overline of Devanagari and closely related scripts confounds optical character recognition software, making it difficult to distinguish letter or syllable breaks. Tamil and other South Indian scripts may be more amenable to machine reading.