I was trying to download the latest snapshot for win32 today but
couldn't run the file because I got an error message about a corrupted
file. D/l-ing again didn't help. Is it my current connection or is the
file really bad?
I'm Gianluca Caporello the main developer of Seats-Tramo, by now, at the bank of Spain we developed a new release of ST which export a C++,Java and Python (and a primitive R) API to ST (Windows and Linux).The API is not documented but it's running from some years on a FAME integration and we have a lot of running examples.
If interested for a better integration in GRETL feel free to contact me or give a look at http://www.bde.es/webbde/es/secciones/servicio/software/econom.html
Here's how to get a crash (with cvs version on linux from Dec 7th):
* open australia data
* do vecm with pau and pus as endo (lag order 4, rank 1), iau and ius as
exogenous, unrestricted, lags 1 to 2.
* modify specification: set all exogenous to restricted
* estimate: -boom-
I just noticed that while it's possible to specify exogenous variables
in the Johansen cointegration test, there are the following issues with
* the exogenous variables aren't reported
* a warning should be printed that the critical values and p-values are
in general only appropriate in the case without exogenous variables
as a change from my flood of bug reports I thought I'd share this
information with you: At the institute where I am working we regularly
receive applications for internships from German (-speaking) students.
Right now it's especially noticeable because the semester break in
February/March is approaching. And what's interesting is that in their
CVs quite many (maybe half?) of the applicants mention that they are
familiar with gretl. That must be because it was taught at their home
university. (And no, they are not applying directly to me, so it's not
because they want to pamper me -- actually I don't think that the
average student looks at the info box and remembers the name of the
So I think gretl is definitely on the rise in the German-speaking region.
have a good weekend,
Hello, our project at launchpad have recently past the stage of manual
import of the template stage and we are able to start translation of
the help system on Launchpad.
Here is the link to the translation in russian:
Here is the link to the template:
I'm currently not on my linux laptop so I will try to generate
.po-files for pt_BR some time later as it was requested. We'll look
what it will be.
---------- Forwarded message ----------
Subject: Translation template import - gretl-help in gretl trunk
Hello Ivan Sopov,
On 2009-12-06 09:27z (5 days 2 hours 52 minutes ago), you uploaded a
translation template for gretl-help in gretl trunk in Launchpad.
The template has now been imported successfully.
The Launchpad team
it seems it's not currently possible to access the results of a
bootstrapped restriction test?
I just did:
stat = $test
p = $pvalue
and stat and p then hold the output of the standard Wald test, which is
misleading IMHO. Instead I would suggest to set $test to NA or something
like that, while $pvalue should give the bootstrapped rejection frequency.
Also, an option to set the number of replications would be useful (e.g.
the following minimal demonstration script:
printf "the period: %s\n", obslabel(i)
yields the following output for me:
Aktuelle Sitzung: 2009-12-09 19:00
? open denmark
Lese Datendatei /usr/local/share/gretl/data/misc/denmark.gdt
Periodizität: 4, max-Beob.: 55,
Liste 5 Variablen auf:
0) const 1) LRM 2) LRY 3) IBO 4) IDE
? loop i=1985:1..1985:4
> printf "the period: %s\n", obslabel(i)
> end loop
loop: i = 1985:2
the period: 1985:1
loop: i = 1985:3
the period: 1985:2
loop: i = 1985:4
the period: 1985:3
loop: i = 1986:1
the period: 1985:4
Anzahl an Iterationen: 4
Notice that gretl's echo/message w.r.t. the loop index seems to be ahead
of the actual loop index by one. But the bug seems harmless, as nothing
substantial is affected.
Or am I missing something?
let's run following script:
setobs 6 2003/01/03 --time-series
series X = normal()
series Y = normal(1,10)
and try to compact data...
But for 5 or 7 day week this option is visible.