GUI points (1)
by Allin Cottrell
At the recent gretl conference in Athens (thanks, Ioannis!) we
discussed potential improvements to some aspects of the gretl GUI.
To avoid overburdening this post I'll confine myself to one point
here, namely access to information about function packages. (I'll get
to some of the other points later.)
As a reminder, in the current release we have a toolbar button labeled
"fx" which supports browsing packages on the local machine or server,
and we have several relevant entries under the Tools menu (access to
the browsers plus other stuff).
One suggestion that got some traction was a "Package" (or "Packages")
top-level menu in the main window (between "Model" and "Help"). The
idea was to move all the package-related menu items from under Tools
to this new menu, to increase the visibility of packages.
I've implemented that in current git and snapshots, but personally I'm
not all that happy with the result. I don't insist on "having my own
way" on this, but I'd like to explain my view and hear what people
think.
First, the "Package" menu has 3 blocks of material: (1) items that
point to the package browsers; (2) items to do with creating/editing a
function package; and (3), if you've edited any packages lately, a
listing of recently opened package files. Note that only block (1),
with its two items, is relevant for "general" users who aren't in the
business of editing packages. All the other stuff, IMO, really belongs
under the File menu because, well, it's all about files. And it looks
wrong to me that it's separated so widely from "File".
So I would favour putting at least the package-editing items (that is,
blocks (2) and (3)) under a new entry "Function packages" in the File
menu (in the section that currently has "Script files", "Session
files" and "Databases"). Access to the browsers could also go there,
but maybe (also?) under Help (possibly combine with the Addons item?).
The File menu surely gets a lot of traffic, so having packages there
would make them somewhat more visible that under Tools. I suppose
there _could_ still be a "Packages" top-level item, but I'm not sure
about using that space for a menu with just two items (the browsers).
I'd also point out that we discussed other ways of making packages
more visible, which I certainly support: sending a regular (say,
monthly) email to the users list with a round-up of new or updated
packages; making packages more prominent on the gretl website; pushing
out info via Facebook and/or Twitter. If we can get those things going
I'm not sure we need a dedicated Package menu.
Allin
8 years
VAR order selection ¿bug?
by Ignacio Diaz-Emparanza
Please run this script:
<hansl>
open denmark.gdt
var 4 LRM LRY IBO IDE --lagselect
var 8 LRM LRY IBO IDE --lagselect
</hansl>
The first four rows should be identical, right? Or am I missing something?
(Gretl built today on Ubuntu 16.04)
--
Firma Arista
*Ignacio Díaz-Emparanza*
ignacio.diaz-emparanza(a)ehu.eus <mailto:ignacio.diaz-emparanza@ehu.eus>
94 6013732
*EKONOMIA APLIKATUA III SAILA (EKONOMETRIA ETA ESTATISTIKA)/
DEPARTAMENTO DE ECONOMÍA APLICADA III (ECONOMETRÍA Y ESTADÍSTICA)
UPV/EHU*
Avda. Lehendakari Aguirre, 83 | 48015 BILBAO
*T.: +34 946013740* | *F.: +34 946013754*
*www.ehu.eus/es/web/ea3* <http://www.ehu.eus/es/web/ea3>
ERNE! Baliteke mezu honen zatiren bat edo mezu osoa legez babestuta
egotea. Mezuak badu bere hartzailea. Okerreko helbidera heldu bada
(helbidea gaizki idatzi, transmisioak huts egin) eman abisu igorleari,
korreo honi erantzunda. Kontuz! Mezua ez bada zuretzat, ez erabili, ez
zabaldu beste inori, ez kopiatu eta ez baliatu.
¡ATENCIÓN! Este mensaje contiene información privilegiada o confidencial
a la que sólo tiene derecho a acceder el destinatario. Si usted lo
recibe por error le agradeceríamos que no hiciera uso de la información
y que se pusiese en contacto con el remitente.
E-mail hau inprimatu baino lehen egiaztatu inprimatzeko beharra.
Antes de imprimir este e-mail piense bien si es necesario hacerlo.
8 years
false local status for fnct. pkg.
by A. T.
Hi,
I've observed that even though I just installed some package from the
server, the package status always remains "Not up to date" -- this holds
in all cases irrespective packages are up to date or not.
I am using latest git on ubuntu.
Best,
Artur
8 years
some issues (some about packages, but not only)
by Sven Schreiber
Hi,
I'm seeing the following things in 2017b on Windows:
1) (not package-related) In the OLS dialog, when clicking "Lags": The
window shows choice for instruments along with regressors, but
instruments doesn't make sense for plain OLS. I'd say that this looked
different "before", where I cannot really define the "before" period,
however.
2) When packaging, the line "tags = C22 C12 ??" in the spec file didn't
provoke an error, maybe it should.
3) Also about packaging, but more complex: The situation is that I have
four public functions, two of which are "menu-only" (and the
"menu-attachment" is specified and works OK, namely in the model
window); but I don't want to have a "gui-main" function, because I did
not want to pick only one of the two.
What happens then is that the only choosable functions after clicking on
the corresponding menu entry are those two that were *NOT* declared as
menu-only. The menu-only functions are basically unaccessible, and I
think that is a bug.
Thanks,
sven
8 years
Gretl-related question on Stackexchange / Crossvalidated
by Sven Schreiber
Hi guys,
following up on the discussion today at the conference, there actually
is at least one question with the gretl tag on CrossValidated (CV) which
hasn't gotten a satisfactory answer (IMHO -- this was asked two months
ago and I missed it because I haven't checked CV in recent months).
Here it is:
https://stats.stackexchange.com/questions/266399/how-to-deal-with-heteros...
It's a perfect opportunity for you people to become acquainted with
Stackexchange / CV!
And to repeat the structure and nomenclature of that site: The whole
meta thing is called StackExchange with around 167 sub-communities. The
communities of interest for us are:
- StackOverflow which is software & programming related
- CrossValidated which is about statistics (including econometrics etc.;
to complete the confusion the URL of that does not contain that name,
but is stats.stackexchange.com)
cheers,
sven
8 years, 1 month
default {} (empty matrix) for matrix args
by Sven Schreiber
Hi again,
not sure if we discussed this before (probably we did): It might be
useful to allow an empty matrix {} as a default argument for a matrix,
instead of "null".
Background/example:
<pseudo-hansl>
function void cm(matrix m[null])
# call to backend function defined elsewhere:
if exists(m)
mybackend(m) # impossible if exists(m)==0
else
mybackend(null)
endif
end function
</pseudo-hansl>
This if-else-endif construction would be unnecessary if we could just
pass around empty matrices as arguments. And instead of the check
exists(m) we could simply use nelem(m) to see what we got.
But I vaguely remember that Allin had some technical objections at the C
level...
thanks,
sven
8 years, 1 month
Mann-Kendall test
by Sven Schreiber
Hi,
I've just read about the non-parametric Mann-Kendall test for a trend
(hence my recent postings about Kendall's tau and ties in a series).
It looked pretty easy, so I'm attaching an attempt at a basic
implementation. Full disclosure: I've used the description from
https://cran.r-project.org/web/packages/trend/vignettes/trend.pdf (only
the basic version in section 1.1).
All you have to do is call "MannKendall(x)", where x is your series of
interest.
I'd be happy about feedback regarding cross-checks and correctness (for
example in the above document in equ. (4) the case S>0 wrongly appears
twice, and I had to make an educated guess where to switch the
signs...), whether it should become a function package and so on.
thanks,
sven
8 years, 1 month
documentation bug about nonparam. correlation (Kendall)
by Sven Schreiber
Hi,
in the documentation for 'corr' it says:
Menu path: /View/Correlation matrix
However, that is not correct for the recently added options --kendall or
--spearman. Those hide under /Tools/Nonparametric Tests/Correlation.
This probably highlights the awkwardness of the View menu. Or perhaps
the nonparam. correlation stuff should be moved from the tools area next
to the parametric correlation. ?
thanks,
sven
8 years, 1 month
how do packages in a model window grab the $model bundle?
by Sven Schreiber
Hi,
in the changelog of 2017b I read:
"Fix bug: ensure that the $model accessor (bundle) is
available to function packages that attach to a model window"
And indeed I have a function here (to be packaged) that takes a
$model-style bundle as an argument. How do I have to specify the
function signature or argument so that it is actually the $model bundle
that is automatically handed over to the package?
I looked briefly in the package guide, but couldn't find an answer (I
think).
thanks,
sven
8 years, 1 month
