Re: [Gretl-users] questions about translation details
Thanks for your feedback!
jack schrieb:
Makes sense to me, although the distinction you draw between "data" and
"sample" looks a bit fuzzy to me: following your logic, the whole
variable-generation machine should go under "sample".
I'd say that depends on whether the commands perform calculations only for the current
sample range
or for the entire dataset range. I confess that I simply don't know which of these is
true, and I
guess both variants have their virtues.
And while we're at it, is the model menu to be left untouched? I
mean, I
wouldn't mind an organisation like
Surely there's room for further opimization within the menus, which I tend to view as
independent
from rearranging items across menus, however.
OLS
-------------
Other least-squares estimators -> IV, Weighted, ...
-------------
Time series
-------------
Qualitative & limited dependent -> Logit, Probit, Tobit
-------------
Count data -> Only Poisson for now
-------------
System -> VARs could go in here
-------------
Other
-------------
no strong opinion from me about this one, although it's probably better than the
status quo... but
VARs are Time Series as well as Systems, so finding the best categories will sometimes be
difficult.
cheers,
sven