RELEASE NOTES 

		spatstat 1.16-3

		12 oct 09

	We thank Bjarke Christensen, Robert Mark, Jorge Mateu,
	Felipe Nunes, Ege Rubak and Alexendre Villers for contributions.

	
OVERVIEW
	Extensions to model-fitting and simulation code.	
	Important bug fixes in 'suffstat' and 'connected'.
	New image metric. 

NEW FUNCTIONS

	deltametric
		Implementation of Baddeley (1992) delta-metric
		(measure of the discrepancy between two binary images)

	distfun
		Returns the distance transform as a function(x,y)
		rather than a pixel image. 

IMPROVEMENTS/ENHANCEMENTS

	simulate.kppm
		Now has arguments 'window' and 'covariates'
		making it possible to simulate a fitted model in a new window
		or with new covariate values.
		[Suggested by Felipe Nunes]

	Kinhom
		The argument 'lambda' can now be a function(x,y)

	ppm
		Now accepts covariates which are function(x,y)
		or a single numerical value. 

	ppm
		Now has an argument 'gcontrol' that can be used to
		control the glm/gam fitting algorithm.

	print.ppm, summary.ppm
		These functions now report whether the fitting algorithm 
		converged (when the ppm object was created).
		Suggested by Ege Rubak.

	quadratcount
	intersect.tess
		These functions now ignore empty tiles.

	suffstat, ppm
		The function 'suffstat' now computes the 
		*edge-corrected* sufficient statistic
		using the same edge correction as the fitted model.
		This should reduce bias in ppm(method="ho").
		
	area.owin
	perimeter 
		These functions now pass the argument w to as.owin
		(suggested by Ege Rubak).

	nncross
		The argument X is now passed to as.ppp()

	[<-.im
		Accelerated.

BUGS

	[1] connected
		The result was incorrect if the input image was not square.
		[Spotted by Bjarke Christensen]
		Fixed.

	[2] suffstat, ppm
		'suffstat' did not perform edge correction.
		Consequently the results of ppm(method="ho") 
		may have been biased in some cases.
		
	[3] rmhcontrol
		Refused to set nrep=1.
		Fixed.

	[4] scanpp
		Crashed when header=TRUE if the column names were not 'x', 'y'.
		[Spotted by Robert Mark.]
		Fixed.

	[5] summary.ppp, duplicated.ppp
		Crashed with message about  'NA in subscript'
		if applied to a marked (not multitype) point pattern
		where the mark values were all identical. 
		Fixed.

	[6] plot.ppp
		crashed if plotting a multitype point pattern
		with more than 52 types. 
		Fixed.