# RELEASE NOTES ## spatstat 1.63-0 ### 23 January 2020 ### OVERVIEW * Compute minimum or maximum nearest-neighbour distance between each pair of types in a multitype point pattern. * Important bug fix in simulations of the multitype hard core process. * Numerous improvements and bug fixes. * Deprecated functions have been removed. * Version nickname: "Trees in space" ### SIGNIFICANT USER-VISIBLE CHANGES * minnndist, maxnndist New argument 'by' makes it possible to find the minimum or maximum nearest neighbour distance between each pair of possible types in a multitype pattern. * beachcolours If 'sealevel' lies outside 'srange', then 'srange' will be extended to include it (without a warning). * split<-.ppp The default for argument 'un' in 'split<-.ppp' now agrees with the default for the same argument in 'split.ppp'. * lineardisc New argument 'add'. Default plotting behaviour has changed. * rmh, rmh.default The printed output of the debugger (invoked by snoop=TRUE) has been improved. * plot.owin New argument 'use.polypath' controls how to plot a filled polygon when it has holes. * plot.profilepl This function has now been documented, and the graphics improved. * erode.owin, dilate.owin These deprecated functions have now been deleted (replaced by erosion.owin and dilation.owin) * delaunay.distance, delaunay.network These deprecated functions have now been deleted (replaced by delaunayDistance, delaunayNetwork) * dirichlet.edges, dirichlet.network, dirichlet.vertices, dirichlet.weights These deprecated functions have now been deleted (replaced by dirichletEdges, dirichletNetwork, dirichletVertices, dirichletWeights) ### BUG FIXES 1. rmh, simulate.ppm, MultiHard Simulated realisations of the multitype hard core model were completely incorrect (the interaction was effectively removed, changing the model into a Poisson process). Fixed. 2. kppm, AIC For kppm models fitted with method='clik2', the resulting value of logLik() was equal to 1/2 of the correct value. This would have affected model comparison using AIC, and model selection using step(). Fixed. 3. hyperframe Did not correctly handle date-time values (columns of class 'Date', etc). Fixed 4. rlpp The resulting pattern was unmarked even when it should have been multitype. Fixed. 5. idw Estimates were zero if 'se=TRUE' and 'power != 2' and 'at="pixels"'. Fixed. 6. model.matrix.mppm Sometimes returned a matrix with the wrong number of rows. Fixed. 7. nncross.ppp Format of output was incorrect if X was an empty pattern. Fixed. 8. rmh, rmh.default For a marked point process, the debugger did not display the marks. (The rmh debugger is invoked by calling rmh with snoop=TRUE). Fixed. 9. pairs.im, pairs.linim The argument 'labels' was sometimes ignored. Fixed. 10. as.im.data.frame Results were incorrect for factor-valued data. Fixed. 11. relrisk.ppp Crashed if there were more than 2 types of points and method = "leastsquares" or "weightedleastsquares". Fixed. 12. as.im.nnfun Crashed when applied to a function generated by nnfun.psp. Fixed. 13. diagnose.ppm Crashed for some models with an error message from 'beachcolours'. Fixed. 14. predict.rho2hat Crashed if one of the original covariates was a function rather than an image. Fixed. 15. lineardisc Crashed in some graphics environments. Fixed. 16. lineardisc Crashed if the network segments had marks. Fixed. 17. rmh Crashed for multitype models if 'nsave' was specified. Fixed. 18. vcov.mppm, simulate.mppm Crashed for multitype models with a hardcore interaction component. Fixed. 19. effectfun Crashed if 'covname' was not the name of a covariate appearing in the model and was not one of the reserved names 'x', 'y', 'marks'. Fixed.