All, for a few years something has been bugging me: Some of the tutorial programs we have aren't quite up to date any more with regard to the tools in the library we use. For example, - step-16 shows multigrid for uniformly refined grids, but we can now do multigrid on adaptively refined grids as well - step-12 shows DG methods, but doing a lot of things by hand that we can now do better and more comfortable with Guido's MeshWorker framework - step-9 shows assembling linear systems with explicit threads when we can do things in a much simpler and more scalable way using the WorkStream framework
Traditionally, what we then did is to just write a new tutorial, or show these techniques en passant in another program (the new step-38, which solves the exact same problem as step-12 using MeshWorker; step-32 which among many other things happens to use WorkStream). The reasoning was somehow that we want to be compatible and not force people to re-read tutorials they've already read. But I'm getting more and more convinced that our current strategy is silly. In particular, it makes people who are new to the library read tutorial programs that just aren't up to date any more. A proliferation of tutorial programs (e.g. step-12 vs step-38) also doesn't help anybody. Does anyone have opinions either way how we should address this problem? Best W. ------------------------------------------------------------------------- Wolfgang Bangerth email: [email protected] www: http://www.math.tamu.edu/~bangerth/ _______________________________________________ dealii mailing list http://poisson.dealii.org/mailman/listinfo/dealii
