[GRASS5] field -> layer

[Radim Blazek]

Michael Barton wrote:
> Radim,
> 
> I'm glad you brought this up. It took me awhile to even figure out what
> "field" meant, given my long association with databases. I agree that the
> name should be changed. But let me ask to make sure exactly how 'field'
> works to make sure I do understand it correctly now.
> 
> A set of vectors (i.e., shapes in a single vector map) can be associated
> with more than one attribute table. 'Field' makes the connection between the
> vectors and each of the the attribute tables. Field1 connects to the default
> table, while field2, field3... connect the same vectors to other tables.
> 
> If this is correct, I suggest using the term connection or link rather than
> layers. Layer can often imply a different set of vector shapes.
> 
> However, if I still don't understand how this works I hope you can set me
> straight.

It is possible to link one geometry to more tables, it is not
the typical case however. Usually, if more fields are used,
the different elements are linked to different tables.
For example, if a landcover map contains forests, lakes and
fields, each centroid is linked only to one table,
using one link. The tables have different structure.
That vector in GRASS is equivalent to for example 3 shapefiles,
3 layers: forests, lakes, field.
Export to simple feature results in 3 layers.
In QGIS (simple feature model), all GRASS features with the
same field are put to one layer.

The name link is not appropriate, because the field can be
a link, but it can be also used in the vector, but there
are no connected tables.

Have you evere played with?: 
http://mpa.itc.it/radim/g51/g51test-12-multi.tar.gz

Radim