On Tue, Apr 29, 2008 at 2:47 AM, Allin Cottrell <cottrell@wfu.edu> wrote:

It'll go either way, just give gretl some guidance as to what you

Yes of course I knew this  way, but as we have talked about consistency recently,  I think that $coeff is supposed to return always a matrix, even 1x1.  That is because with multiple regressors $coeff is the matrix of coefficient, whereas with only one regressor it would the scalar of coefficient. Moreover there is at least one case when I cannot know if the regressors will be either single or multiple: that is when I have a user-defined list of regressors.
Anyway this is just my opinion about a really negligible issue.