- numerical (in)stability
just the first example might be perfect for SCBC2 or SCBC6: http://www.flow3d.com/cfd-101/cfd-101-computational-stability-I.html
might be an interesting choice, depending on how the FORTRAN went: http://www.ibiblio.org/pub/languages/fortran/ch4-7.html
I see the Dendy text mentions that things should be numerically stable, p. 180
—
Just an interesting resource overall:
http://spiff.rit.edu/classes/phys317/phys317.html
—
Perhaps useful for SCBC2:
http://en.wikipedia.org/wiki/Floating_point#History
—
When Jens was here, he mentioned that Software Carpentry had written two papers wrt (with respect to) best practices:
http://software-carpentry.org/blog/2012/10/best-practices-for-scientific-computing.html → http://arxiv.org/pdf/1210.0530v4 though I can't say I agree with 1.4 (missing from the paper, I note) or 7.2 (5.2 in the paper).
http://www.flow3d.com/cfd-101/cfd-101-computational-stability-I.html haven't had a look yet
—
Good example: “Computational simulation can greatly enhance scientific understanding by allowing the investigation of simulations that may be difficult or impossible to investigate by theoretical, observational, or experimental means alone. In astrophysics, for example, the detailed behavior of two colliding black holes is too complicated to determine theoretically and impossible to observe directly or duplicate in the laboratory. To simulate it computationally, however, requires only an appropriate mathematical representation (in this case Einstein's equations of general relativity), an algorithm for solving these equations numerically, and a sufficiently large computer on which to implement the algorithm.” -Heath, page 2
—