Matlab has fft and ifft already to fast calculate discrete Fourier transform and its inverse. However they are binary built-in functions, hence their codes are not visible, which is not helpful to someones who want to understand how FFT works. For teaching and learning purposes, I think we should also have a pair of their readable counterparts, written and stored in plain text (the so-called the Matlab m-functions), which we can step through during their executions. I searched Mathworks file exchange center but did not find any one had contributed such m-functions. Therefore I hope that the m-functions that I am going to contribute will be welcome to the Matlab user community.

To avoid name conflicting, I name my pair of functions sfft and sifft. Using letter s as a prefix is learnt from Professor Gilbert Strang in MIT, who presented slu as readable counterpart to the Matlab built-in function lu for lower and upper triangular matrix factorization I do not know why he chose s at the first place. Perhaps it was derived from the word script, something written in text.