Page locked

This page is currently locked for editing by another user. You have to wait until this user finishes editing or the lock expires.

Currently locked by: 216.73.216.147,10.2.117.91
Lock expires at: 2025/07/04 21:12 (12 min)

This page is read only. You can view the source, but not change it. Ask your administrator if you think this is wrong.

=== List of potential benefitial changes which would break old input files ===

  - Renaming of elements. As suggested in the meeting in Prague (April 2013) we need to rethink the naming conventions for elements. Names should specify, in order: problem type, geometric type, order. The names need not be perfect (it gets tricky with specialized structural elements) but should reveal enough about the implementation to make them easy to understand.
  - No default dofs from the domain anymore: The "domain **domaintype**" should be removed. The number of spatial dimensions should instead be set directly by the input field "nsd #" in the domain record.
  - Dirichlet boundary conditions should be applied to sets instead of directly on dofs. The standard dirichlet b.c. should also take a vector argument for the prescribed values, to go along with "dofs", e.g. "dofs 3 1 2 3 value 3 0. 0. 1." which denotes that: D_u = 0., D_v = 1, D_w = 1.
  - With no more default dofs from the domain type, controlling dof-type in hangingnode, slavenode or rigidarmnode should be done by supplying the optional "dofidmask" field.
  - LoadTimeFunction should be renamed to simply "TimeFunction" as they are used generally.
  - Nonlinearstatic should use the normal newton solver by default.
  - NodalLoad should use the "dofs" input field to determine which components are directly loaded.