Hi,
In response to Caroline Lambert's question, my approach in Candis
(one of the grandparents of netcdf) has been to use reduction operations
to make a large file into a small one -- i. e., to cut out particular
slices using Candis filters, to average over dimensions, etc. Also,
since Candis has a very easy way to make graphical output, I rarely
find myself looking at lists of numbers anymore. If the netcdf filters
ever get implemented and released, this approach would work well with
netcdf also.
Dave Raymond
