I think letting users add their own boundary elements might be too complex(?) so some kind of control through keyword arguments or extra Xsection classes might be the way to go. But this too could get pretty complex given all the options here (finite domain, fixed head, ghb, flux, or boundaries between domains with fixed heads or resistances in specified layers). Something to think about :).
What I was thinking is: We can make, for example, a River (line-sink) element where the user specifies the layers it is in, and in the other layers applies a continuity condition. That is all that is necessary. So it is an extended element, but the element itself and the specification by the user is easy. What remains to be decided is how to specify this for inhomogeneity boundaries. Maybe we can catch that for users: When an element is placed on a boundary, we replace the boundary (inhomogeneity) element by this new element. And your nudging code can be removed. Let's think about it some more. Definitely something for a new Issue.
Originally posted by @mbakker7 in #107 (comment)
What I was thinking is: We can make, for example, a River (line-sink) element where the user specifies the layers it is in, and in the other layers applies a continuity condition. That is all that is necessary. So it is an extended element, but the element itself and the specification by the user is easy. What remains to be decided is how to specify this for inhomogeneity boundaries. Maybe we can catch that for users: When an element is placed on a boundary, we replace the boundary (inhomogeneity) element by this new element. And your nudging code can be removed. Let's think about it some more. Definitely something for a new Issue.
Originally posted by @mbakker7 in #107 (comment)