Dear members Many authors of papers or speakers at seminars will talk about a pioneering Statistical method and say in passing that while no standard package can handle it at present, 'this can easily be implemented on the computer' - but this clearly applies only to those who have the requisite knowledge and training. There don't seem to be many books that offer this type of training, I have seen for example books teaching Fortran 90 which are oriented towards engineering - but nothing for Statistical computing. Can anyone recommend good books on this subject that are written specifically for Statisticians who might wish to implement a new method on the computer, but which no package (as yet) can handle? Regards Miland Joshi (Mr.) Department of Epidemiology and Health Sciences University of Manchester Medical School %%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%