I've been meaning to write a full response to your comments on
the overall gui design (you have thought about this carefully
and certainly deserve a response), but I haven't had time yet.
At present I only have time for a brief response:
Thanks Professor Cottrell, yes I have been thinking about this
carefully. I see that both you and Jack disagree with my suggestions
for now and I understand and respect this perfectly. Still, I cannot
help but make a few additional comments to clarify myself:
Firstly, I didn't wish to imply that I was hoping to see these
suggestions implemented immediately for the 1.6.0 version. This is a
quite rather radical change and a lot of testing and other users'
feedback are required to implement it.
That being said, I honestly think that these changes can really be a
desirable and feasible direction for the future versions. I came to
this conclusion after talking with a few people who are users of other
programs and not very familiar with gretl. For example, my wife is
using eviews for her dissertation and she says that the only reason
she would even consider switching to gretl is the icon view. She has a
huge dataset and constantly running tens of different regressions and
dealing with many many graphs. She says it is really frustrating for
her to keep track of everything she has been doing and there is no way
she can easily organize all her work in eviews. (still, she cannot
switch because she is doing some dynamic panel stuff). The same thing
goes with me. I am using gretl for my dissertation and the icon view
is a lifesaver for me to get organized. I have so many icons lying
around.
The current setting is very efficient for classroom use, where the
students open a sample data file and easily see the frequency
distribution or run several tests by highlighting a variable from the
variable list. However, with the proven accuracy and constantly
increasing number of tools and functions, gretl is headed to become a
fullfledged econometric analysis environment, not just a "library" of
functions.
Imagine a user planing to start a rather comprehensive analysis with
gretl. The user is encouraged to first define a new project or a
session since his or her work is likely to take more than an hour or
two. Even if it is a small project, it is a good practice to archive
all the work in the end so working in sessions should be encouraged
always. That is why I propose putting session related items at the top
of the menu. A .gretl file is what .xls is for excel or .nb is for
mathematica. (By the way, I see that maybe there is no need for a "new
session" item since other programs have it when it is possible to work
with multiple projects. Can having multiple sessions be something
desirable in gretl?)
Similarly, the icon view becomes more important as the size of the
project increases. Once the users starts a new project and adds a data
file or a null data, all the subsequent analysis should be savable as
icons. Maybe there will be multiple sampling of the data. The rules
for each sample can be saved as an icon for example. Think of a truly
object oriented environment and each object as an icon. Matrix
functionality can probably be transfered to the GUI with the help of
icons as well. (Maybe a matrix palette or toolbar?) I agree with Jack
that this is a difficult task but it should be possible. This is where
intuition and lots of thinking, trying and refining comes in.
Now, both Allin and Jack says that variable view conveys very
important information and functionality. I understand this and propose
that the "variable window" can be open by default. I created a mockup
screen, which is attached to this message. I think it looks nice and
existing functionalty is not sacrificed. Of course it can be better
than this.
Finally, I disagree with the idea that some things are better left for
the console in order to encourage users to use the command line more.
IMHO that defeats the purpose of using gretl. On the other hand, I
perfectly understand that all these changes require resources and
there are other priorities such as implementing more functionality.
Here is hoping that all these will be possible in the future with the
additional help by new volunteers. As Allin says gretl is increasingly
becoming a community project. So I think it is good to have a clear
direction for the future of the program now. By the way, I think
somebody has already registered the
www.gretl.org. Entering address
forwards to
www.gretl.com (which is also taken). It just displays a
blank page with "GRETL" written at the top.
Sincerely,
Talha
--
A positive attitude may not solve all your problems, but it will annoy
enough people to make it worth the effort. - Herm Allbright
(1876-1944)
--