On Thu, 2 Jul 2015, Riccardo (Jack) Lucchetti wrote:
On Wed, 1 Jul 2015, Allin Cottrell wrote:
> Good points. I have updated the top-level gretl webpage accordingly.
> Those who maintain translations of the website please note. See
> where the relevant (English) file is index_pat.html.
I would like to propose a few additions/corrections:
* GARCH -> "wide variety of univariate GARCH-type models"
* should we also mention ordere logit and probit, multinomial logit etc?
* "Heckit" is a popular name, but a little frowned upon by some; I quite like
it as a house term, but on the webpage I would prefer "sample selection
* Under panel-data estimators, also include probit, ivpanel; use the "GMM"
* Say something about robust/nonparametric stuff
* "Command loop structure" is obsolete. I'd rather have something more on
hansl instead. A link to the hansl primer wouldn't hurt, either.
* I'd probably add a few examples of what you can find among contributed
function packages. Actually, this could somehow be merged with the
"Contributing to Gretl" section.
More good points. I'll follow these suggestions, except that I won't
attempt to do much with the function packages section for now -- that
seems to me a bigger issue.
And while we're on the topic, we really ought to have a forum.
experiment more or less died a peaceful death, but nowadays I guess that
people expect to be able to ask questions to "the community" through a forum
rather than mailing lists. Of course, looking after a forum website is a
hefty job (it's important to filter spamming etc), so I don't expect anybody
to volunteer in the next 5 minuts, but let's just say it'd be nice if someone
stepped up ;)
If we were going to do that, it would make sense to use some existing
machinery such as stackexchange. Any volunteers to find out what
exactly it would involve to set up a gretl "site" there?