On Fri, 23 Jan 2009, Allin Cottrell wrote:
 On Fri, 23 Jan 2009, Riccardo (Jack) Lucchetti wrote:
> Since few people seem to read the Changelog, here's a "press release"
> summarising the main changes in the new version. Comments welcome.
 Very nice.  I'll make sure that goes up on sourcefore. 
Thanks. This one should be somewhat better (research assistence by "cvs 
diff" is gratefully acknowledged).
-------------------------------------------------------------------
Version 1.8.0 of gretl is out. Here is a list of the major
changes with respect to 1.7.9. As usual, you can download the
latest version of gretl from
http://sourceforge.net/project/platformdownload.php?group_id=36234
USER EXPERIENCE
---------------
* Switch to compact tabular form for presenting model statistics
* Graphical User interface:
     - Add model menu item: "Modify model", under the Edit menu,
       provides a clone of the model specification
     - graph controller: new facility to add a line, defined
       via a formula, to an existing graph;
     - line-color selector: extend to 6 colors, and distinguish
       between setting the "palette" and setting ad hoc colors for
       a specific graph
     - Boxplots: hand the production of these over to gnuplot
     - relocate user-defined scalars from main window to the
       icon view window;
     - many other small improvements
* Data filters
     - add importer for SPSS .sav files
     - CSV importer: try to handle time-series data in reverse
       chronological order
* Many, many bug fixes
COMMANDS AND FUNCTIONS
----------------------
* New commands:
     - "intreg": interval (grouped data) regression
     - "modprint": makes it easy for script writers to format
       estimation output
* Improved commands:
     - "fcast": allow variables for the startobs and endobs
       arguments; add "rolling" k-step-ahead option
     - "gnuplot": now works in loops (batch mode only)
     - "loop": now more robust, more efficient and more general
     - "mle" and "nls": New option --numeric to force the use of
       numerical derivatives
     - "ols": add --jackknife option for robust standard errors and
       --anova for printing the ANOVA table
     - "ols" and "panel" (fixed effects): add option to calculate
       p-value for the Durbin-Watson statistic
     - "restrict": now available for nls, mle and gmm
     - "tsls": now offers the --liml and --gmm options as
        alternatives to two-stage least squares
* Modified commands:
     - var: the matrices returned by the accessors $sigma and $vcv
       for VAR systems now have a degrees of freedom correction
* Obsoleted commands: "rhodiff" and "hccm"
* New functions:
     - Add accessors $Fstat and $chisq for the overall F-statistic
       or chi-square test from the last model
     - boxcox(): Box-Cox transformation
     - mcovg(): matrix covariogram
     - mpols(): enables multiple-precision OLS for user-defined
       matrices
     - toepsolv(): solves a Toeplitz system of linear equations
     - urcpval(): gives user access to James MacKinnon's
       p-values for unit root and cointegration test statistics
       (tau)
* Improved functions:
     - mlag(): now handles multiple lags using a vector as the
       second argument
     - pdf(): Speed-ups when the function is applied to a series
       argument
-------------------------------------------------------------------
Riccardo (Jack) Lucchetti
Dipartimento di Economia
Università Politecnica delle Marche
r.lucchetti(a)univpm.it
http://www.econ.univpm.it/lucchetti