Re: [Gretl-devel] Signalling
On Tue, 8 Jan 2013, Logan Kelly wrote:
Hello all,
As one of the new users you would like to recruit, I would like to add
some feed back on this topic. First, let me season my comments by saying
that I went from skeptic to fanatic in about a month. Gretl is simply
one of the best packages available (pay or free).
Nie to hear :)
If the goal is to make Gretl more popular, then more time needs to be
focussed on usability. Even if that comes at the short-term expense of
developing new capabilities.
First, the GUI, whether you stick to the menus or are a coder, is your
first impression of the software. And the GUI really understates this
"grader" of Gretl. I would love to see a more integrated GUI with docked
windows like STATA or Matlab.
This has been discussed several times over the years. People from a
windows background (of course the majority) tend to like applications
which show as a big "container" of "subitems". Maybe we should revisit
this topic once more.
Second, the documentation. While the user guid is a good start, parts
are incomplete and I have thus found myself asking simple question on
the email forum. (Thank you for graciously putting up with me as I am
learning). Also, why is there a function reference and a command
reference? I am continuously looking thing up twice because I do not
know if I need a function or a command. Simply combining these documents
in some way that makes sense would be helpful.
Yeah, this is a bit of a sore point. Trouble is, writing docs takes time,
an awful lot of time. IMO, if Allin and I and anyone who's willing to help
could find the time to finally have the hansl manual done, the situation
would improve drastically, in that you'd have:
1) a short user's guide, for beginners and gui-only people
2) the hansl manual, for scripters
3) a separate "Econometric methods" guide, for those who want to know
_exactly_ how gretl does things without perusing the C source (a bit à la
"Methods and Formulas" sections in the Stata docs)
4) The Reference manual
and this arrangement would probably be less confusing.
What I would love to see is that
1. The user guid and help files be put in HTML so that like STATA I can
Google a method and quickly find a help file, e.g. googling "Gretl ols"
returns a help page on running ols in Gretl.
Nice idea. The source is in xml already, so this could be done pretty
quickly (at least in thoery).
2. The email forum is good, but there are several forum type q&a
projects that would (i) be easier for new users to search and access and
(ii) show up in search engine results. This would require manually
coping content to the new forum, but that is something new comers like
me could contribute to the community. (I know this is a topic that has
been raised before, but I think it is very worth consideration).
The wiki experiment hasn't been a success, to put it mildly. We may try.
3. More video tutorials of basic GUI functionality. This would (i)
make
adoption in stats courses easer and (ii) increase search engine
visibility. (Note: I am planning on making a series of tutorial for an
MBA course I teach.)
Great!
Finally, I think citing the use of Gretl should be at least strongly
suggested in exchange for using the package, similar to Ox.
This is another point what hasn't been resolved for years: we need to find
a way to cite gretl in such a way that is consistent and shows up on the
major engines, such as Econpapers/IDEAS, and IMO we need this badly.
-------------------------------------------------------
Riccardo (Jack) Lucchetti
Dipartimento di Scienze Economiche e Sociali (DiSES)
Università Politecnica delle Marche
(formerly known as Università di Ancona)
r.lucchetti(a)univpm.it
http://www2.econ.univpm.it/servizi/hpp/lucchetti
-------------------------------------------------------