RELEASE NOTES 

		spatstat 1.9-2

	 	5 June 2006

OVERVIEW

	Major improvements to the summary functions (Kest, Gest, Fest, ...)
	Improvements to model-fitting code.
	
NEW FUNCTIONS

	persp.im
		Perspective plot of a pixel image.

		This previously undocumented function now has a 
		help page. Display has also been improved.
		Images are now displayed in the correct x:y aspect ratio.
		There is a new argument 'colmap' for easy control of the 
		colour map.

	profilepl
		Computes the profile pseudolikelihood of the 
		irregular parameters in a point process model, 
		and finds the best-fitting values.

IMPROVEMENTS/ENHANCEMENTS


	Kest, Kmulti, Kcross, Kdot, Kinhom, Kcross.inhom, Kdot.inhom 
	Gmulti, Gcross, Gdot, Jmulti, Jcross, Jdot, 
	Iest, pcf.ppp, markcorr	
		
		These functions have been accelerated by performing
		more of the basic computations in C.
		The new algorithms will handle point patterns of 
		several thousand points, for all edge corrections.
		
	Fest, Gest, Gmulti, Gcross, Gdot, Iest, Jest, Jmulti, Jcross, Jdot, 
	Kest, Kmulti, Kcross, Kdot, Kinhom, Kcross.inhom, Kdot.inhom,
	Kest.fft, pcf.ppp, allstats, alltypes, envelope, markcorr:

		These functions no longer require that the r values
		span the range of interpoint distances being measured.
		
		These functions have a new rule for determining the
		default range of r values. 

	ppm
		Computation has been accelerated 
		for models with pairwise interaction.

	update.ppm
		New argument `use.internal' makes it possible
		to avoid re-evaluating the arguments to ppm.

		This also makes it possible to repair the damage 
		to a ppm object caused by dumping to a file and reloading it.

	plot.ppm
	predict.ppm
	residuals.ppm
	fitted.ppm
	diagnose.ppm
	qqplot.ppm
		These functions now automatically detect the damage
		to a ppm object caused by dumping to a file and reloading it.

	pairdist.ppp
	crossdist.ppp
		New argument 'periodic' for periodic edge correction

	pairdist.default
	crossdist.default
		New argument 'period' for periodic edge correction

	km.rs
		Now works correctly even if the breakpoints do not span
		the data.

	kaplan.meier
	reduced.sample
		New argument 'upperobs' to kaplan.meier
		and new argument 'uppercen' to reduced.sample
		allow these functions to be applied to histograms even if
		the breakpoints do not span the data.


BUGS

	[1] border corrections and reduced sample estimates

		The border correction estimates in Kinhom
		and the reduced sample estimate in Fest, Gest, Jest
		were possibly incorrect if a non-default value of the
		argument 'r' was given by the user. This occurred
		if the range of r values did not span the range of distances
		(pairwise distances for Kinhom, empty space distances for Fest,
		nearest neighbour distances for Gest). 

		Fixed.

	[2] eval.im 
		the result did not always have the correct format
		when one of the images was factor-valued.
		Fixed.

	[3] quantile.im
		Did not handle NA's properly. Fixed.


	[4] ppm objects
		If a ppm object was saved to a file using
		dump() or save(), and subsequently reloaded using
		source() or load(), this damaged the internal
		format of the object, and many operations would crash.

		Such damage is now detected automatically by the 
		functions plot.ppm, predict.ppm, residuals.ppm, 
		fitted.ppm, eem, diagnose.ppm and qqplot.ppm.

		This damage can now be repaired by calling
			update(object, use.internal=TRUE)