Hi all,
due to a recent discussion in the user list
https://gretlml.univpm.it/hyperkitty/list/gretl-users@gretlml.univpm.it/t...
Allin and I have begun coding a slight modification to the --robust option
to ols that enables the usage of that option with "gappy" time series
data.
A few minutes ago, I pushed a commit to the git master branch that makes
the new behaviour conditional on the existence of an environment variable
named "HAC_ALLOW_MISSING". This is a technique we've been using quite
often to test new things.
In fact, maybe _too_ often. The combined powers of grep and sed allowed me
to compile a list of all the environment variables that play a role of
some kind; although some are perfectly legitimate (eg HOME, LANG), some
are clearly there because of experiments we did in the past. I guess we
ought to clean up at least some of them.
Of course, Allin and I could see to this without bothering the community,
but this may be a good opportunity for aspiring developers to take a peek
at the source code and see what's in there, so ideas and contributions
from the community are very welcome.
The list is:
ARELLANO_ASY
CLI_NO_PLOTS
DFGLS_NO_PVALUE
FORCE_GFN_PATH
GMM_NORMALIZE_2
GRETL_CONFIG_FILE
GRETL_FORCE_SEED
GRETL_HOME
GRETL_INCLUDE_PATH
GRETL_MATRIX_DEBUG
GRETL_MATRIX_PRINT6
GRETL_MP_BITS
GRETL_MPI_HOSTS
GRETL_NO_RLIB
GRETL_OLD_EV
GRETL_PLUGIN_PATH
GRETL_PNG_GRAPH_FONT
GRETL_REAL_SVD
GRETL_USE_OPENMP
GRETL_WORKDIR
GRETL_WWW_VERBOSE
HAC_ALLOW_MISSING
H_EXTRA
HOME
IMPORT_DEBUG
JSONGETB_OLD
KALMAN_DEBUG
KALMAN_DEJONG
KALMAN_EXACT
KALMAN_TRACE
KALMAN_UNIVAR
LANG
MINGW_PREFIX
OMP_NUM_THREADS
PATH
REPLACE_NAIVE
REPLACE_USE_BTREE
R_HOME
SHELL
SHELLDEBUG
VAR_USE_QR
Have fun!
-------------------------------------------------------
Riccardo (Jack) Lucchetti
Dipartimento di Scienze Economiche e Sociali (DiSES)
Università Politecnica delle Marche
(formerly known as Università di Ancona)
r.lucchetti(a)univpm.it
http://www2.econ.univpm.it/servizi/hpp/lucchetti
-------------------------------------------------------