I've made some changes today in response to a request that Talha a
few days ago, and I'd like to get people's opinions on whether
this change should stay.
That is, should gretl dialogs (other than simple "instant
response" dialogs such as errors and warnings) be marked as
"staying on top" of other windows? [Note: it can't be guaranteed
that this will be respected by all window managers, but there's a
mechanism in GTK to request this.]
Talha's argument was that this firmly prompts users to respond,
rather than "losing" the dialog under other windows. I think that
has some validity, but I can also see a case for saying that a
program that pushes its dialogs in the user's face is being too
bossy. What do people think?
One reason I ask is that I've thought up till now that we couldn't
do this for dialogs that have a Help button, since (if anyone ever
reads Help!) the dialog would obscure its own help text. I've
thought of a way to get around this problem, but before I try
coding I'd like to hear if we want to go with "stay on top"
dialogs at all. (That is, other than error dialogs, which do need
to force a response.)
On the topic of GUI changes, here are a couple of ideas. Some could be
served with small changes (improving icon view tooltips), some others might
require a lot of work; as I have no idea of the amount of work involved
please excuse any unreasonable idea I might've unknowingly made of the kind
"I don't like gtk, please reimplement everything on .net". Of course there
are resource tradeoffs and these 'projects', as any, may not be the smartest
use of them possible even if you find them reasonable...
* Add a "define list" command to the "add" menu in the main window.
* The user-created scripts could be listed on a separate tab in the
"scripts index" dialog (file, script files, practice file...) for better
* Don't auto-close that dialog when opening a script file.
* Add a "run script" button to that dialog's icon bar.
* Replace the "new script" button on the main window with a "scripts index
* (easy version of the above) A few more than 4 items in the scrips and
sessions file lists (under files menu) would be nice.
* Allow more than one model table.
* Add a "save to model table..." command in the save menu of the model
result windows for all model tables that apply (that is, that share the same
* The icon view could be a little better as model names aren't descriptive
(e.g. name: "model 1", tooltip text: "ols 1 9 34 5", and I might not
remember what 1 9 34 and 5 stand for, so I begin opening models to see what
have I've done previously). It can get messy if you, e.g., save a model,
then a chart and then another model, as icons are not arranged by type. Some
graph icons don't show a tooltip indicating what was graphed. So maybe a
list with columns may be better: type (table, graph, etc), name, description
(what's charted or model specs).
Thanks for reading,
Great work on the gretl software, it is extremely user-friendly compared to other econometrics programs. I have two questions I was hoping someone could answer.
1) if my model was something like the following: Mt = B1 + B2*Rt + B3*Yt + B4*Mt-1 + B5*Mt-2 + u
- how would I perform a durbin wu hausman test for the hypothesis that my variable Rt is exogenous, if I use Rt-1 and Rt-2 as additional instruments
- also, how would I estimate the above equation by generalized instrumental variables, treating Rt as endogenous, and Rt-1 and Rt-2 as additional instruments?
2) For a non-linear model such as: y = B1*x1 + B2*x2 + B3*x3^g + u
-how can I use a succession of Gauss Newton Regressions to obtain the NLS estimates of all the parameters and their standard errors?
Thank you so much for any help that can be provided.
Hello GRETL list members :
I am looking for a webcourse that provides GRETL code for the analysis
of time series. Can you please provide with some links to those
If you can not be a mentor neither you can be a professor nor you can
be an advisor!
101 Ag. Adm. Building, Louisiana State University A&M
Baton Rouge, LA USA 70803-5604
Tel. mobile - (225) 4057851 Email: cgarci8(a)tigers.lsu.edu
On Tue, Dec 1, 2009 at 11:17 PM, Allin Cottrell <cottrell(a)wfu.edu> wrote:
> On Tue, 1 Dec 2009, Talha Yalta wrote:
>> If I am not mistaken, about three years ago we've had a discussion
>> regarding a reorganization of the gretl menus, which resulted in a
>> much improved user interface. Since gretl keeps getting more and more
>> functionality and new menu items, this may be a good time to do this
>> again. As a result:
>> 1)- I support Allin's proposal to merge "Data", "View" and "Sample" to
>> 2 headings.
> Hmm, I did float that idea, but I'm not sure how to do it.
I think there is definitely room for improvement in the current gretl
menu structure. Here is a humble proposal by yours truly:
PROPOSAL FOR THE ADVANCEMENT OF THE GRETL MENU AND WINDOWS STRUCTURE
FOR A BETTER ACCOMODATION OF THE INCREASED CAPACITY AND FUNCTIONALITY
SINCE THE VERSION 1.6.0
by A. Talha Yalta
December 2 2009
In July 2006, the gretl-users mailing list has seen a lengthy
discussion on how to restructure the gretl menus, which were perceived
at the time by the users and the developers to have become too
cluttered due to the increased functionality by the version 1.5.1.
This discussion can be accessed from the archives at
Thanks to the lead developer Allin Cottrell's efforts, the "menu
re-structuring" discussion has resulted in significant improvements of
the gretl graphical user interface (GUI) in the subsequently released
Over the last three years, the gretl econometrics package has
continued to advanced rapidly and significantly and with the new
functionality came more and more menu objects. It is our understanding
that by the current version 1.8.6, gretl is getting close to a point
where there are too many menus and menu items and the placement of
some of the objects are not necessarily optimal. The objective of this
proposal is to present our views on how to achieve a slimmer and
tidier gretl menu arrangement that can also accomodate the new
functionality with further development in the future.
The proposeal is tructured as follows: We simply list the proposed
changes below. This is followed by a brief conclusion.
1)- The sample menu will be a sub-menu of the data menu. To accomplish
this, it first needs to be slimmed-down as follows:
a)- The count of missing values will be added to the status window.
b)- The "Define based on dummy..." item will be removed since
this can be easily (and even better) accomplished using the "Restrict
based on criterion..." item and with simple criterions such as
dummyname=1 or dummyname=0.
c)- "Set missing value code..." item will be sent to the
preferences menu and will be renamed as "Default missing value code
for datasets." When the user changes this setting, gretl will ask
whether the changes will be applied to the current data set. (The
missing value code can be considered as dataset dependent, however, we
see the above-described handling of similar issues in other programs
as well. For example, in openoffice the default language settings for
documents are in the preferences menu although this is also document
With 3 less items, the sample sub-menu will be composed of the following:
Restore full range
Set a range...
Restrict based on a criterion...
Resample with replacement...
Drop obs with missing values
As an alternative to having a sub-menu, the above functionality can be
offered in a single sample window accessed from the data menu. The
sample window can show the status and and at the same time let the
user make changes and see the results immediately in the updated
status information in the same window.
2)- The "View" menu will be renamed as "Window" and will be composed of:
(The list of all the currently open windows)
Refresh windows (if this item is really necessary)
3)- The "Data" window will:
a)- Acquire the items abondoned by the former "View" menu
b)- Have a new sub-menu entitled "Modify dataset," which contains
the below items that are not accessed very frequently:
Change data structure...
4)- The "Variable" menu needs to be slimmed down a bit. It will be
composed only of those items regarding a single selected variable
(Please note that it becomes ghosted when there are more than one
a)- The "Set missing value code" is already sent to the
preferences as discussed in (1c) above.
b)- The "Edit attributes" item will be moved to the "Data" menu
next to the "Dataset info" and "Observation markers..." items.
c)- The "Define new variable..." item will be removed. I
understand that it is a very common task, however, it is a fundamental
GUI design principle that there has to be exactly one menu item for a
given task. Cut-copy-paste is very common too but we do not see this
trio in more than one menus in a word processor or a spreadsheet
program. Imagine what would happen if we had more than one set of the
same buttons in an elevator or a plane cockpit.
d)- The renamed "Find a variable..." item will be somewhere else
(maybe in the tools menu?).
5)- The "File" menu is in pretty good shape except for the "Working
directory..." item, which clearly belongs to the Preferences window.
6)- Regarding the "Preferences" sub-menu:
a)- The "Preferences" sub-menu will subsume the "Fixed font..."
and the "Menu font..." items and become one "Preferences" window.
b)- It will also include a much needed "Check status" item which
will list whether various functionality such as LaTeX support, Gnu
Multiple Precision support, MPFR support, FFTW3 support, X-12-ARIMA
support, TRAMO/SEATS support etc. are working in the current
7)- The recently added "windows" icon (compass earlier) in the various
windows will be removed in favor of a simple interface with less
clutter. The restructured menus described above allows an easy reach
to any window from the main gretl window via the new "Window" menu and
the recently added right click menu item "Main window" now makes it
very easy to go to the main window.
8)- As discussed earlier in the gretl-users list, the title of the
various windows originating from different models should reflect the
original model. For example, instead of having "gretl:ANOVA" for all
ANOVA windows, the title will be "gretl: ANOVA (Model 8)." This is
also necessary to prevent the proposed "Window" menu having various
windows items with the exact same name causing confusion.
9)- Various open dialog windows waiting input from the user will stay
above other windows.
10)- We believe that a gretl menu re-structuring process presents a
very good opportunity to improve the LaTeX output options in the
"Summary statistics" and the "Correlation matrix" windows as well.
In conclusion, as a long-time user and a contributor of gretl, it is
our understanding that the above described modifications will result
in a significantly improved user interface that is not only relatively
simpler but also more intuitive. We are sure that further comments and
suggestions by the other users and developers will help us achive the
excellent GUI expected from a high-quality software such as gretl.
A. Talha Yalta
“Remember not only to say the right thing in the right place, but far
more difficult still, to leave unsaid the wrong thing at the tempting
moment.” - Benjamin Franklin (1706-1790)