These are the NIST statistical reference datasets for linear
regression analysis.  To run a given test, do (e.g.)

./test Pontius

This will generate an output file named client.out for the
dataset Pontius.dat.  The results may be compared with the definitive
multiple precision figures given in the comments in Pontius.hdr.

You should find the reults agree with the definitive ones, up to
the precision given by gretl/eslclient, except for Filip.dat,
which is rejected by gretl as too highly multicollinear.
