
13.dec 2004
===========

- if devi is not given rmsi computation is skipped
- segmentation needs to be checked whether all fixes were incorporated
- reading if z is attribute leads to swapping for large data sets
- crossvalidation needs improved efficiency
- add crossvalidation to description
- remove the dmax option
- change error message if segment is smaller than grid cell (in data.c)

old notes
=========
- it says ...identical points instead of
           ....point closer than given dmin (or whatever we have in s.surf.tps)

- check whether masking works

- test variable smoothing
