new and updated features in eRm 0.9-6:

- infit and outfit mean-square statistics added to functions
personfit() and itemfit()

- new method: plotPImap()
plots a person-item map (cf. Bond & Fox, 2007), i.e., a map
of locations of item (and threshold) parameters, and the
distribution of person parameters

- new options in plotGOF():
'conf': draws confidence ellipses for item parameters,
optionally single items can be chosen interactively
'ctrline': draws confidence bands along the diagonal line
based on the se's of the item parameters (cf. Wright &
Stone, 1999) plotGOF() now displays difficulty
parameters(see below)

- unification of output for item parameter estimates:
all plotting functions and the function thresholds() display
item difficulty parameters. all numeric output (i.e. from
model objects) gives item easiness parameters.

- new and enhanced options in plotICC():
'empirical': empirical ICCs can now be smoothed using
various smoothers (such as Tukey, loess, or kernel)
'ask': interactive turning over of plots can be switched
off (only useful if automated figure export is in effect,
e.g., when using Sweave)

- simulation module for 0-1 response matrices:
sim.rasch() for Rasch homogeneous data,
sim.2pl() for 2-PL data,
sim.xdim() for unidimensionality violation, and
sim.locdep() for locally dependent item responses.

- package vignette "eRmvig" added