Comment on ´pulses satisfying the Nyquist criterion´ (and reply)

For original paper see vol.28, no.1, p.50 (1992). The Nyquist pulse shape proposed for increased immunity to symbol timing error was presented nearly 25 years ago as the optimum pulse for any small timing offset and a prescribed pulse bandwidth (see IEEE Trans., Vol CoM-16, no.2, p.337). The problem was posed as a quadratic minimisation problem and solved by variational calculus techniques.