1:03 p.m.
I used XSIM, so I'm enthused to jump on this excellent thread!
James.Callahan wrote on July 20, 2005 9:12pm:
Allin Cottrell wrote on 07/14/2005:
<<snip>>
> What I have here is more of an attitude than a definite plan,
but
> I'd like for gretl to be able to do much of what Eviews can, but
> better(!).
I don't have Eviews, so what sorts of things?
<<snip>>
What I would like to see is a macroeconomic forecasting environment
that
is a superset of gretl.
My inspiration/itch is my experience at Chase Econometrics in the early
1980s and what we could and were trying to do on mainframe virtual
machines.
The environment had several elements:
1. XSIM - regression, multi-equation simulation/modeling, time-series
database, scripting language, report writer (similar to TROLL)
Alas, XSIM is now used by the PLA to simplify input of chinese ideograms!
Back then, some features were
() access many huge databases without loading data, or even directories, that weren't
needed.
() a simulation model compiler that output machine code
() mix multiperiod data, including daily (Sun-Sat, Mon-Sat, Mon-Fri or once a week).
() direct access to a huge international securities data collection,
() write your own FORTRAN, PASCAL routines ("plug-in")
() hierarchical, slice-n-dice multi-dimensional data cube
() a scripting language that was getting better integrated with applications,
even as the product died.
Derived from TROLL(MIT), it fostered AREMOS at WEFA, Hyperion, etc.
But it was based on tty "CLI" technology,
could not link multiple simulation models or mixed periodicity automatically,
or solve rational expectations,
lacked great high-level visual debugging diagnostics for bad models and data,
or ... (plus cost A LOT).
Is there a name for the integration of output charts as graphical input media?
E.g., in Meta-stock or TC2000, every point in the stock price chart is a hotspot.
If the mouse hovers, a little yellow box pops up with local information in it,
whether it's just the date on the x-axis or the exact value from the y-axis
or the name of the point or line.
If you right-click you get a context-menu of other things to see or do.
Can GnuPlot do this for Octave and R?
I also would like to be able to drill down/back from a result to its input.
Say, when a variable value is an aggregate or computed,
Visualising a structural model as a graph between nodes.
I fantasize animating a structural model with a little cursor and dotted line
starting at a point on an x-axis, going up to a fitted line, then over to the y-axis...
...THEN jumping to another plot where z = g(y), etc, filling a plasma panel
with a structural model unfolding as a mouse pokes around -- all visual,
no reading and re-typing anything.
2. Chase Econometric Macroeconomic forecasts (US & International)
3. Chase Econometric Macroeconomic databases (US & International)
<<snip>>
PYTHON interfaces to a number of GUI tools including Tkinter and WXWidgets
(wxpython).
http://wiki.wxpython.org/
What is the difference between GTK and wxWidgets? I.e., how did you choose?
Professor Roy Fair at Yale has his Fairmodel which is includes a US
and a
MultiCountry (MC) macroeconomic models and he regularly publishes
forecasts produced with the models. he also makes all of the data used to
estimate the model available.
http://fairmodel.econ.yale.edu/main2.htm
Could Eclipse be the development environment that integrates these tools
for economic modeling, rather than Java programming?
<<snip>>
HERE'S THE PAYOFF:
A professor teaching Freshman macroeconomics could have students do
cookbook excerises such as estimating a consumption function or doing a
predetermined macroeconomic simulation, "just type the commands in the book."
A professor teaching a Junior year second course in macroeconomics could
have students do more open-ended exercises.
<<snip>>
In the best case scenario this might even lead to
economically literate, model based discussions of macroeconomic policy
options....
FUTHER PAYOFFS:
Once you have reliable source of macroeconomic forecasts one can build
industry and regional models.
<<snip>>
Another fantasy is to shadow a debate on TV, or a presentation on the
McNeill-Lehrer Newshour, in real time with data, equations, charts in
some sort of Wiki. Principals could cite their own websites, instead
of Ross Perot holding up a pie chart. The debate might refer to input
data or equations or statistical techniques, and viewers could play with
alternatives or follow along,
and maybe even submit questions embodied in alternate models,
all because GRETL is a free world standard.
Yipes, I do run on...
Would people care to offer a more structured wish-list (wiki?),
citing where the feature would be used and/or
how big a job it is or how many starving kids it would feed or
whether it is an innovation or competitive derivation?
I'm looking for a small, summer project...