I once worked in an IT department of a genetics laboratory, which consistently publishes high impact papers in journals like Science and Nature.
While I was there I got to see some of the R/Python/Java code that had been created for all kinds of studies that had taken place there. It was some of the worst code I have ever seen.
But I do agree, the complete methodology including the code should be shared.
> Psychophysics Toolbox Version 3 (PTB-3) is a free set of Matlab and GNU Octave functions for vision and neuroscience research
Scientists almost never want to share their code and it makes me sad. Why wouldn't they want to make their paper more easily reproducible?