On Fri, 13 Oct 2006, Sven Schreiber wrote:
Allin Cottrell schrieb:
> One other point: when you estimate a model via the gui (single equation
> or multiple) it does _not_ become the "last model", for data retrieval
> purposes, so far as the console is concerned. This is perhaps debatable,
> but it's by design. The idea was that things could get confusing.
> Suppose a user has one "thread" going in the console (estimate, retrieve
> values, calculate), and then, "on the side", estimates a model via
> point-and-click. Then where should, say, $uhat be drawn from in the
> next console command?
> At present the answer is, from the last model estimated via the console.
So this goes both ways, does it? I mean that also a new console/script
estimate does not affect the model which is used for menu commands?
Yes. When you have a model window open, the window is "shadowed"
by an in-memory copy of the model in question, so when you save
data from a given model window you should always get data from
that particular model, regardless of any other models that may
have been estimated more recently.
I can see the point, although personally I use (or would like to
use) gretl differently.
I can see a case for having the behaviour that you expected (i.e.
estimating a model via the gui updates the "last model" as seen
from the console). That wouldn't be difficult to do.
Btw, is there a kind of system to track
enhancement requests like these? I might get around sometime to
contribute some documentation, but not right now. And without a
bugtracker this may be forgotten.
There is such a mechanism on sourceforge. I haven't used it much
to date, but it's there.
Anyway, I guess my stumbling point was just the fact that
matrices aren't yet supported via the gui. It would be nice if
that could be added, but I understand/agree that that doesn't
have a high priority.
There's some exploratory code in place, not activated yet. I'll
get around to doing some more on that at some point.