On Wed, 18 Jan 2006, jack wrote:
The good news is that adding GMM is probably doable in a few weeks
(at least 1-step GMM, but probably 2-step can be done in that
timeframe too); the bad news is that neither Allin (I presume) nor
I are able at the moment to tackle this, because we're rather busy
with some new exciting things...
I might just add a hint or two at what's going on (I agree, this is
exciting).
* We're adding a general facility for defining and manipulating
matrices. This is fairly well advanced but needs more testing. At
present we have most of the standard matrix operators and functions
implemented, using a syntax much like ox. People using gretl CVS
can read about this in the new manual chapter file, matrices.tex.
* We're also adding a general facility for "farming out" tasks from
gretl to other programs that offer functionality not (yet) available
in gretl. We already do that to some extent with x12arima and
tramo/seats, but Jack has developed a more general, elegant and
efficient mechanism that will allow gretl to exchange binary
information with external programs. Ox is the primary target at
present, but R is also a good candidate.
In combination, these facilities will greatly increase the
opportunities for people to write functions (in gretl script, not C)
that extend gretl's capabilites.
I'm also working (though this is at an early stage) on a gui
interface to such "add-on functions". The idea is you could
download these from some repository, then open up a gui window
showing what functions are available (and who wrote them, and what
their purpose is), with the option of loading any functions of
interest into gretl's workspace, where they could be used in scripts
or at the gretl console. Perhaps these add-on functions could even
attach themselves at appropriate positions in the gretl menu
structure, and hence become directly accessible via point-and-click.
Allin Cottrell