Hi Barton,
> Dan Stanger wrote matrix exponential code that uses Putzer's ode
> based method, but I can't find it in share.
I can't find it either.
> In short, we have two methods, but each package has good stuff the
> other doesn't. The intersection is 'mat_function' (diag) and 'matrixfun'
> (linearalgebra). This isn't the only or the last such conflict.
>
> Do we want a scheme such as 'diag:mat_function' and
> 'linearalgebra:mat_function'? One CAS handles packages
> this way.
Well, I certainly do want a package scheme.
Maybe it's time to reopen that discussion.
We should still try to cut redundant code, but given that
the two implementations mat_function / matrixfun are
different in an interesting way, it is reasonable to put them
in separate packages.
best,
Robert