On Tue, 22 Aug 2006, Riccardo Jack Lucchetti wrote:
> Thanks. Well then, we're not out of line with all other
> programs. (I suppose ox is really a programming language rather
> than an "econometrics package" as such, and in programming
> languages 0-base indexing is certainly the norm.)
And even that, not always: IIRC, FORTRAN is 1-based too, and
so should be gauss and matlab. I think base 1 is ok, it's one
less thing for people without a C background to worry about if
they want to port their procedures into gretl scripts.
OK, agreed, we'll go with 1-based.
Attached: a little script that does the first, easy part of
the Verbeek GMM example.
Allin.