@tomleslie 

thankyou , S[] gives x

study brent method found old post

why

 tol_act := EPS * abs(b) + 0/3;

0/3 - is typo or not

@acer 

perfect just what want, ok for [1,1] = coeff z(x,y)

new version of H is good to, can use either

thanks lot

@acer 

works good thanks

is possible for partials like

f = f0(x,y) + f1(x,y)*diff(z(x,y),x) + f2(x,y)*diff(z(x,y),y) + f3(x,y)*diff(z(x,y),x,x) + f4(x,y)*diff(z(x,y),x,y) + f5(x,y)*diff(z(x,y),y,y)  ...

maybe matrix like

m[1,1] = 0
m[1,2] = coeff(f,diff(z(x,y),x)
m[1,3] = coeff(f,diff(z(x,y),x,x)
m[1,4] = coeff(f,diff(z(x,y),x,x,x) ....
m[2,1] = coeff(f,diff(z(x,y),y)
m[2,2] = coeff(f,diff(z(x,y),y,x) = coeff(f,diff(z(x,y),x,y)
m[2,3] = coeff(f,diff(z(x,y),y,x,x) = coeff(f,diff(z(x,y),x,x,y) ....
m[3,1] = coeff(f,diff(z(x,y),y,y) ....

again thanks

@Tamour 

easy, go here https://webstore.maplesoft.com/index.aspx, pay money, get working copy

you want help, but cheat maple business - very bad, why should users help pirate

@acer

package work good, questions

why linalg not LinearAlgebra

possible use without float[8] input - maybe want lu/solve of matrix [1, 2; 3, 4] - exact not float - maybe option to input type -  exact/float, what about high digits say Digits=100 input matrix - works to 100 accuracy or must change module, maybe module deal all digits case automatic by data check - is possible or no

thx for great package

why you keep ask for people do your homework, think for yourself

@Carl Love thanks for other way, true i confuse order

@tomleslie what i want thanks, M maybe has term like x1*sin(x2)*cos(x3) so need what you give

@Rouben Rostamian   good idea functions that return not only u, but also u_x, u_xx, and u_xxx.

@Rouben Rostamian  how to get data u(x,t), data for chose t value, data for chose x value thx

@Rouben Rostamian  works now thx

good app, doesnt work on maple 2017, how to fix thx