Strange behavior of Fenics sample algorithm for Navier Stokes equations

Not in the webpage, but you did in this post

In an effort to contribute and improve this beautiful project, I suggest that the definition of p^{n-1/2} and p^{n-3/2} and the fact that p is solved at steps 1/2, 3/2, 5/2, … while the velocity is solved at steps 1, 2, 3, … should be made crystal clear in the webpage.

I will revise my code accordingly and see whether those odd jumps in sign go away!