Remove doc/modules/todo.h

[class4kayaker]

The sole commit removes doc/modules/todo.h(#1459), but should not be merged until all of the issues below have either been noted to have been handled either by being closed, be ongoing (not certain best way to specify/handle this, but a github issue not owned by a maintainer is probably not suitable), or have a linked github issue.

I have noted my understanding of the state of each of the raised issues, but would like confirmations before opening any issues.

Contents of doc/modules/todo.h

Extensions we'd like to implement

• Mark compositional fields as to their meaning. This may include:
  1. concentration: $0 \leq c \leq 1$
  2. binary: $c \in \{0,1\}$
  3. positive: $c \geq 0$
  4. level set: $c \in \mathbb{R}$
  5. arbitrary: $c \in \mathbb{R}$

  Depending on the meaning of a field, we may want to use different
  artificial viscosity approaches, compression, etc. We may also want to
  project back into the feasible set when passing compositional values to
  material models.

  ---

  Incomplete, (have method, not type, may be sufficient) Mark meaning of composition field #1526

• Provide an option to store not only the last snapshot but many

  ---

  Complete

• Rewrite the tracer code

  ---

  Completed

• Move the adiabatic heating term (the last one in the
  temperature equation) to the left hand side and make it implicit

  ---

  Complete

• Make snapshotting as atomic as possible so that a job that is
  terminated while creating a snapshot doesn't get us in trouble

  ---

  Should be evaluated by a maintainer, but probably complete

• Implement "open boundary conditions" as described in "Using open
  sidewalls for modelling self-consistent lithosphere subduction
  dynamics", M. V. Chertova, T. Geenen, A. van den Berg, and W.
  Spakman see discussion on the aspect mailing list "[aspect-devel]
  strain rate corner values and velocity boundary conditions"

  ---

  Ambiguous, do not see a plugin, but may be implemented or have available equivalent

• The nonlinear solvers are at best lightly tested. Do this
  more systemtically.

  ---

  Ambiguous, may be ongoing project

• The same can be said of the compressible solvers

  ---

  Ambiguous, may be ongoing project

• Self gravity (Ian Rose has something initial)

  ---

  No obvious plugin

• We need a scheme to verify that plugins are compiled against the
  same version of ASPECT that they are running under.

  ---

  Incomplete?

• More benchmarks. One possibility is the benchmark by King that has
  been described (apart from the original paper by Scott) in the
  thesis at
  https://www10.informatik.uni-erlangen.de/Research/Projects/terraneo/docs/schlag-thesis-2014.pdf
  starting at page 22.

  ---

  Particular one mentioned is complete? (King2dcompressible)

  May be ongoing project

Bugs we know of

• Name the "table" material model better than "table"

  ---

  No table material model

• The temperature equation must contain adiabatic terms in the
  compressible case.

  ---

  Uncertain status, not sure where to check

• Re-write the geometry section of the manual since information on the
  boundary components is no longer there

  ---

  Probably complete, but could use confirmation from a maintainer

Features missing tests

• tests for the various refinement criteria

  ---

  Probably completed, but need to confirm

• write a testcase for friction heating (Poiseulle flow) and adiabatic
  heating (constant downward velocity)

  ---

  Probably completed (friction_heating and adiabatic_heating)

• write tests for the depth averaging functions (test adaptive
  refinement, averaged composition, non-zero average/sinking velocity)

  ---

  Probably completed

Other issues to address

• We should collect a list of publications and posters created with
  the help of Aspect

  ---

  Ongoing project, any opened issue should probably be owned by a maintainer

• Unify readme.html in aspect/ and from the webpage again

  ---

  Complete?

• Use extrapolated solution for grid refinement

  ---

  Not currently done (Is this desired?)

• Make initial conditions class say whether it implements a
  temperature or a temperature perturbation

  ---

  Currently not done(delayed with composition init for 2.0?)

[gassmoeller]

Thanks for going through the list. I have checked everything I found solved, not important, and opened new issues for the remaining topics. This should pave the way for removing todo.html. Do you want to open a pull request?

[gassmoeller]

Ah, my bad, this is already the pull request. I will wait for the tester and then merge.

[gassmoeller]

/run-tests