Re: [Gretl-devel] we need defense and/or marketing
Am 11.11.2015 um 08:58 schrieb Riccardo (Jack) Lucchetti:
On Tue, 10 Nov 2015, oleg_komashko(a)ukr.net wrote:
> Dear all, A citation The following is a list of free open-source
> software. We do not teach the use of these programs in our courses.
> We teach using software that you may encounter is the workplace.
> Support for these packages is limited
I actually agree with the main point of that citation: if you educate
students who are not (all) going to become academics, you also have an
obligation to provide them with marketable human capital. And (for
example because of the network effects that Jack mentioned) gretl is
just not very widespread yet. So in the past I've tried to not use gretl
in class exclusively.
Not a big deal IMO. The attitude "you get what you pay for" has a long
tradition. I delivered a seminar at a big research institution a couple
of months ago (gretl-unrelated) and at lunch guy told me about his
mistrust about free software because "if you pay for a piece of
software, you have an implicit guarantee that it works properly"; to
which I replied that he probably uses a web browser he never paid for,
while OTOH I never heard of anybody sueing successfully StataCorp or QMS
for bugs in Stata or Eviews (or MS for bugs in Excel, for that matter).
The guy had a puzzled look, probably he had never considered things from
that angle.
Again I agree mostly; however, I once found a bug in Eviews (which was
subsequently corrected with a patch by the company). In contrast, I have
long lost count of the number of bugs I found in gretl. If it hadn't
been for the exceptional attitude of Allin in acknowledging and fixing
these bugs quickly I would have given up on gretl long ago. I think it
is an objective fact that gretl has been much more buggy. And to be
honest, the numerical accuracy of Eviews has been much improved over the
years I think.
So the point is: I think the public is right that they want to have a
big entity behind the software. I don't really think they necessarily
mean "company", a big community is also ok; see the case of R and Python
which have been adopted widely in recent years.
It is a fundamental problem that the gretl community is too small.
However, I think the right actions to correct that have been taken
already, most notably the function package area.
...
Again, advocacy is ok, but only works up to a point. Time is on our
side.
I also think that apart from continuing the work there is not much else
to be done.
cheers,
sven