On Mon, Mar 18, 2013 at 3:01 PM, Kirk, Benjamin (JSC-EG311) <[email protected]> wrote: > On Mar 17, 2013, at 5:51 PM, Roy Stogner <[email protected]> wrote: > >>> >>> What polynomial is used to create the hierarchic basis? >> >> Hopefully Ben or John will chime in on this one. In 1D the >> p-hierarchic property is enough to determine the basis up to scaling, >> but in 2D and 3D it gets more complicated, and even in 1D I don't know >> how the scaling was chosen. > > svn log is not much help here, which goes all the way back to r3, nor is my > memory. > > All I have to offer are the cryptic comments int the code: > > // All even-terms have the same form. > // (xi^p - 1.)/p! > case 2: > returnval = (xi*xi - 1.)/2.; > break; > case 4: > returnval = (pow<4>(xi) - 1.)/24.; > break; > case 6: > returnval = (pow<6>(xi) - 1.)/720.; > break; > > // All odd-terms have the same form. > // (xi^p - xi)/p! > case 3: > returnval = (xi*xi*xi - xi)/6.; > break;
I think these were added by Steffen Peterson or Daniel Dreyer. They both worked directly in our CVS repo in the CFDLab before libmesh was put on sourceforge, but it appears that all those logs were unfortunately lost when we ran cvs2svn. All the even ones look like bubble functions, the odd ones aren't... -- John ------------------------------------------------------------------------------ Everyone hates slow websites. So do we. Make your web apps faster with AppDynamics Download AppDynamics Lite for free today: http://p.sf.net/sfu/appdyn_d2d_mar _______________________________________________ Libmesh-users mailing list [email protected] https://lists.sourceforge.net/lists/listinfo/libmesh-users
