Do there exist "frameworks" as to how computational scientific experiments claim validity?
Like "scientific method for computed science"?
$\begingroup$
$\endgroup$
3
-
$\begingroup$ Do you mean like a Karl Popper-style "philosophy of computational science"? $\endgroup$– user14717Commented Feb 23, 2019 at 17:43
-
$\begingroup$ @user14717 Something that outlines good scientific practices for computational science. $\endgroup$– mavaviljCommented Feb 23, 2019 at 17:49
-
$\begingroup$ Is John Claerbout's reproducible computational science roughly what you want? $\endgroup$– user14717Commented Feb 23, 2019 at 18:09
Add a comment
|
1 Answer
$\begingroup$
$\endgroup$
There is a lot of research on techniques for verification and validation of codes that use numerical approximations to solve partial differential equations arising in fluid dynamics and other areas of engineering and science. The focus here is on verifying that the code solves the discretization of the problem correctly and that the discretization is doing an adequate job of approximating the PDE so that the resulting solution will be good enough for engineering use. See for example:
W. L. Oberkampf and C. J. Roy. Verification and Validation in Scientific Computing 1st Edition. Cambridge University Press, 2010.
P. J. Roache. Verification and Validation in Computational Science and Engineering. 1998.
answered Feb 24, 2019 at 0:02