Roger.Bivand at nhh.no
Wed Apr 23 08:00:11 EDT 2003
On Wed, 23 Apr 2003, Radim Blazek wrote:
> On Wednesday 23 April 2003 12:16 pm, Markus Neteler wrote:
> > There was some time ago a discussion on NULLs
> > http://grass.itc.it/pipermail/grass5/2001-June/000307.html
> > Maybe interesting in this context.
> I like this reply to that question in user list:
> Yes, why we need separate file for NULL values? How easy
> (code maintenance) and fast (faster PC not needed) would it be
> just to have one value defined as NULL.
We _have_ been here before: legacy GRASS rasters do not know about NULL,
so an additional structure is needed to implement NULL as an extra layer,
so that they can still be used in current GRASS. There are - as we know -
lots of modules, and many have typically not used the current API, so they
would break (this may no longer be correct, but was correct), if we change
the low-level representation. Of course, it would be logical, but it looks
to me like a 5.3 thing, not a 5.1 thing especially since both machines and
disks seem to get faster (but user patience gets shorter?).
Economic Geography Section, Department of Economics, Norwegian School of
Economics and Business Administration, Breiviksveien 40, N-5045 Bergen,
Norway. voice: +47 55 95 93 55; fax +47 55 95 93 93
e-mail: Roger.Bivand at nhh.no
More information about the grass-dev