Allin Cottrell schrieb:
Jack and I know JMulti, and have used it for testing some of
VECM functionality. It's a nice program.
> May be we are interested in some type of link or collaboration?
We should refer to it in the gretl docs along with R and Octave, as
related free software. I'm not averse to the idea of some sort of
collaboration, but I'm not sure what form it might take.
Yes it should be mentioned. Actually, I know Markus (the developer of
JMulti) because we're both from Berlin. (It's not that Berlin is so
small, but when you deal with time series econometrics and are about the
same age, then that happens.)
At this point, I think the most useful starting point for a
collaboration would be to make the datafiles convertible. I guess that's
already a non-trivial task.
Btw, in principle you can call Jmulti's routines from "anywhere" (as
Java classes). With Markus's help we verified that for example you can
use the Python language for that (using its Java implementation Jython).
I can't speak for Markus, but my impression is that in some sense JMulti
is a finished product in bugfix mode. But it's open source and in
principle can be extended, but I think gretl's new function interface is
more flexible in that respect. (not performance-wise, I guess)
So much for that from me,