TriangularInterpolation and SetPrecision bug

*To*: mathgroup at smc.vnet.net*Subject*: [mg23633] TriangularInterpolation and SetPrecision bug*From*: Sergio Perez <Mathematiker at yahoo.com>*Date*: Thu, 25 May 2000 01:00:44 -0400 (EDT)*Sender*: owner-wri-mathgroup at wolfram.com

I use TriangularInterpolation from the package ExtendGraphics to create an interpolating function [salt] over a set of 3d data. When I apply the function to a matrix of 3d coordinates[data] inside the area interpolated some values are read as outside the domain of the interpolating function (lets say, salt[data[[i,j,1]],data[[i,j,2]]] for some [i,j] give {Indeterminate,Indeterminate}). Nevertheless the interpolating function applied to the explicit numeric values itself of [data[[i,j,1]],data[[i,j,2]]], lets say salt[12.8,6] gives the right outcome (Lets say 1.96). I have tried to change the format of the argument of the interpolating function (lets say [data[[i,j,1]],data[[i,j,2]]] ) to allow the function to read them as "different" numbers using: 1) [IntegerPart [data[[i,j,1]]]+ FractionalPart[data[[i,j,1]]],IntegerPart [data[[i,j,2]]]+ FractionalPart[data[[i,j,2]]]] or 2) [N[data[[i,j,1]]],N[data[[i,j,2]]]] but this does not work as the problem is still presented. 3) Things get worse using SetPrecision as [SetPrecision[data[[i,j,1]],3],SetPrecision[data[[i,j,2]],3]] because the calculations are not only long but the outcomes are completely crazy. If someone has tackled this kind of handicap please let me know.