Is it correct to define Lagrange linear basis functions on p2 isoparametric triangular element.
We know the subparametric elememts whose shape functions are high order polynomials (n>=2), but the edges of the element are straight.
I wonder whether the 'inverse' definition is correct, more precisely:
Define low order polynomial basis functions on high order isoparametric elements.
I tried to solve Poisson equation on unit circle, and in order to approximate the curvilinear boundary, I wanted to use quadratic (cubic) isoparametric element with utmost one curvilinear edge (on the boundary).