no longer interpolate

[gdmcbain]

In #692 , 57d8e29 shows in

scikit-fem/docs/examples/ex13.py

Line 66 in 57d8e29

current[port] = asm(port_flux, fbasis, u=u)

that it's no longer necessary to interpolate a field passed in as a coefficient to the assembly of a form. This issue is a reminder to propagate this to the rest of docs/examples, as promised, once #692 is completed and merged.

[kinnala]

There was recently some confusion regarding the use of other finite elements as material data inside the forms. Let's say that u was defined using fbasis_p3 instead of fbasis, then explicit call to fbasis_p3.interpolate is required because assemble would automatically call fbasis.interpolate(u) and not fbasis_p3.interpolate(u).

The most common use case involves using same fbasis but I wonder if we start dropping interpolate in the examples, will it cause more confusion?

[gdmcbain]

Yes, I see what you mean. I don't know…. I don't think I'll get confused but it's hard to say for new users and I don't have any at the moment, maybe next year. I think I'll withdraw this for now unless anyone else has a firmer opinion. Thanks.