RELEASE NOTES 

		spatstat 1.11-2

		12 March 2007
	
OVERVIEW

	New dataset. Several new functions and improvements.
	Changes to the handling of illegal points in a point pattern.

NEW DATASET

	anemones
		Kooijman's beadlet anemones dataset
		from Upton and Fingleton (1985) 

NEW DEMONSTRATION

	demo(data)
		Type demo(data) to display all the datasets
		installed with spatstat.

NEW FUNCTIONS
	
	ks.test.ppm
		Kolmogorov-Smirnov test of goodness-of-fit
		for a Poisson point process model.

	[<-.im
		A method for "[<-" for pixel images has been implemented.
		This makes it possible to change the pixel values in 
		a subset of an image. See help("[<-.im")

	ewcdf
		Weighted version of 'ecdf' (empirical cumulative distribution
		function)

	clickpoly
		Interactively draw a polygonal window

	bounding.box.xy
	convexhull.xy
		Convenience functions for finding the smallest 
		rectangle or convex polygon containing a set of points.

IMPROVEMENTS/ENHANCEMENTS


	[.im
		The subset method for pixel images now accepts
		a logical-valued pixel image as the subset index 'i' in x[i]

	Kmeasure
		Now has argument 'varcov' allowing anisotropic smoothing.

	quadrat.test
		This function is now generic, with methods for 
		point patterns (quadrat.test.ppp) and 
		for point process models (quadrat.test.ppm). 
		The argument 'fit' is no longer needed or used.

	bounding.box
		Now accepts arguments of class "ppp" and "im" as well as "owin"

	as.im
		Now has argument 'dimyx' to control pixel array dimensions.
		Behaviour improved.

	ppm
	quadscheme
	default.dummy
	spatstat.options
		The minimum allowable number of dummy points 
		in a quadrature scheme is now controlled by
		spatstat.options("ndummy.min").

	ripras
		New argument 'shape' 

	print.ppm
		Output is slightly neater.


CHECKING POINT PATTERNS

	A ppp object contains (x,y) coordinates of points
	together with a spatial window. It is essential that
	the points in a ppp object lie inside the window.

	In spatstat 1.11-0 and 1.11-1, when a new dataset was created
	using ppp(), the system checked whether the (x,y) points 
	all lie inside the window. If there were any such `illegal' points
	a fatal error was generated.

	For the convenience of users dealing with messy data,
	in spatstat 1.11-2 we have changed the following functions:

	ppp
		If any points are found to lie outside the window,
		instead of causing a fatal error, 
		the `illegal' points are now removed from the point pattern,
		and the remaining `legal' points are returned as a ppp object.

		A warning message reports how many points were illegal.

		The `illegal' points are still accessible, as an attribute
		called "rejects" of the ppp object. 

	plot.ppp, print.ppp
		If a ppp object has an attribute called "rejects"
		then this object is also plotted/printed.

	as.ppp
		as.ppp removes the "rejects" attribute if there is one.

	all other functions
		All other functions ignore the "rejects" attribute,
		and use only the `legal' points.
		
		
CHANGED DATASETS

	copper
		The components of the copper dataset named 'points' and 'lines'
		have been renamed 'Points' and 'Lines' to avoid
		name conflicts when the copper dataset is attached.


BUGS

	[1] persp.im
		Crashed in R 2.4.1 with message "illegal 'z' limits"
		if the image was constant.
		Fixed.

	[2] plot methods
		Some of the plot functions crashed with the message
		     'Internal error: function "points" not found"
		This occurred if a dataset called 'points' existed.
		Fixed.

	[3] affine
		For the case where the window is a rectangle and the
		transformation matrix is diagonal,
		  (i) code was incorrect; 
		    the shift operation was not applied to the window.
		  (ii) an error occurred if there were negative values
		   in the diagonal of the transformation matrix.
		Fixed.