Jim Lemon wrote (in response to a question from Alberto Monteiro): > I once had to do this to generate "random dot stereograms" for > perception experiments. One easy way is to build from one corner or > edge (almost all these are rectangular, you can just clip out the > figure you want later). The user defines a "hit" area for each > successive point that contains the minimum and maximum allowable > distances from the two nearest points (typically a small square). > Points can be generated with two uniform random numbers having a mean > of the distance to the center of the "hit" area and a range spanning > the "hit" square. It was easy to churn out constrained random dot > patterns on the fly with this method.
This appears to be: (a) closely related to the ideas of ``partially ordered Markov models'' developed by Noel Cressie et al (see, e.g. Cressie, N., Zhu, J., Baddeley, A. J., and Nair, M. G. Directed Markov point processes as limits of partially ordered Markov models. Methodology and Computing in Applied Probability, 2, 5-21). (b) re-invention of the wheel. As I pointed out in a previous posting to the list there are a brazillion ways of generating point patterns with interpoint inhibition, readily available in R. cheers, Rolf Turner [EMAIL PROTECTED] ______________________________________________ R-help@stat.math.ethz.ch mailing list https://stat.ethz.ch/mailman/listinfo/r-help PLEASE do read the posting guide http://www.R-project.org/posting-guide.html and provide commented, minimal, self-contained, reproducible code.