RELEASE NOTES 

		spatstat 1.6-1

	 	14 March 2005


OVERVIEW

	In Version 1.6 we will expand the functionality of the package
	by adding
		- new model-fitting techniques
		- improvements to simulation algorithms
		- residuals and diagnostic plots for fitted models
 
	We start in release 1.6-1 with an internal reorganisation
	of the Metropolis-Hastings algorithm implementation. 
	This should be compatible with any code that worked in 1.5-10.
	The algorithm will run slightly faster.

COMPATIBILITY

	There are many internal changes, but they
	should not be visible to the user.
	
	The defaults for `control' in rmh, rmh.default and rmh.ppm
	have changed.

NEW FUNCTIONS

	rmhmodel
	rmhstart
	rmhcontrol
		These functions now pre-process the arguments 
		'model', 'start' and 'control' of rmh.

		You do not need to invoke them explicitly,
		so existing code will still work.

		On the other hand, invoking them explicitly
		and saving the results will reduce computation time 
		when calling rmh repetitively to generate many realisations
		of the same model.

	reach
		Calculates the interaction distance of a fitted point process
		model or an interpoint interaction structure

	rescue.rectangle
	expand.owin
		Two very simple functions for fiddling with windows.


IMPROVEMENTS/ENHANCEMENTS

	1. the Metropolis-Hastings algorithm runs slightly faster.

	2. rmh.ppm now works with expanded windows (control$expand > 1).

	3. rmh.ppm and rmh.default now allow expanded windows
	to have periodic boundary conditions (control$periodic=TRUE).

	4. In rmh.ppm and rmh.default, 
	control$expand can be a window object instead of an expansion factor.

	5. The defaults	for `control' in rmh, rmh.default and rmh.ppm
	are now consistent.

	6. is.subset.owin gives better answers in marginal cases.

	7. as.im and as.owin are now more accurate.
	For example as.owin(as.im(W)) is now the same as W, for any window W.
	
BUGS

	[1] rmh.ppm did not handle the case of conditioning on n
	when the model is Poisson. This has been fixed.

	[2] There were bugs in example(rmh.default) that did not
	cause errors. They have been eradicated.