Re: [Help-gsl] You are looking for an expert in numerics?

[Brian Gough]

Roland Sonnenschein writes:
 > Hello, I have seen your request on "Help wanted" . Perhaps I can
 > help your project?  At numerics of ODEs, monte carlo simulation and
 > approximation of transcendental functions I am quite good, I
 > suppose.  I speak C/C++, Fortran77, bash, gawk and different
 > flavours of assembler fluently.  My native language is German, so I
 > will never consider to write a manual in English . Nevertheless I
 > can still do proofreading focussing contents.

Hello,

Thanks for your email.

Maybe you can help us fix a problem with the implicit ODE solvers --
it is mentioned in the BUGS file (BUG number 5).  Some of the error
terms are not implemented correctly.    

For example in ode-initval/rk[24]imp.c and ode-initval/gear[12].c the
error terms are calculated as O(h^n) without the correct coefficient
or dimensionality.

If you could look up the algorithms in the literature and find the
correct error terms that would be very useful.

best regards,

-- 
Brian Gough