[Qgis-developer] About digitizing with GRASS Plugin

[m roy]

Thanks Radim now it's more clear.

Il 19/10/2015 17.46, Radim Blazek ha scritto:
> On Mon, Oct 19, 2015 at 3:05 PM, m roy <royroge at outlook.com> wrote:
>> Il 19/10/2015 12.41, Radim Blazek ha scritto:
>>> It is not possible to add in QGIS a layer without a type. So when a
>>> new empty layer is created, a type must be chosen. The type selected
>>> by user is only used for the layer added to canvas/legend, the created
>>> vector map is empty.
>> Ok, I see, could it be better to disable point and line tools
>> while working with polygons and so on? Just an idea.
> You can digitize other types. Say that you add a new point layer, you
> can digitize points, lines and areas, but QGIS layer (in non editing
> mode) can only display points. Once you close editing, other types
> should appear in browser and you can add them to canvas. So you can
> digitize (almost) everything at once, but QGIS lets you visualize only
> one layer/geometry type when the layer is not edited.
>
>>>> (2) *Attribute table editing*:
>>>> not able to figure out how it works, i managed
>>>> to edit single feature using the "Identify Features"
>>> Changing the values in table cells (when editing) does not work?
>> Opening the attribute table i get a lot of records with NULL values
>> I can imagine that those are for boudaries and the filled ones are
>> centroid records (?)
> Yes, all primitives are represented as features (that is the only way
> to talk with QGIS). Boundaries usually do not have attributes but may
> have. For example, if a road runs around a forest, you can use the
> boundary of the forest to form forest area and represent the road at
> the same time. In such case, the road will have probably different
> attribute fields. Currently only one layer of the same map can be
> edited at time. Multi layer vectors are not yet perfectly supported
> however.
>
> I see the problem with field calculator.
>
>>> Regarding the symbology of boundaries, I am convinced that the
>>> distinction between area/area
>>> and area/isle used in GRASS editor is useless while distinguishing of
>>> boundaries with topological error is essential for effective editing.
>>> Radim
>> Well actually in GRASS you simply have different colors based
>> on the topology of boundaries, but I understand now,
>> given the QGIS layer limitations you explained at point (1),
>> that this may not be possible because QGIS layer does not
>> have any information about boundaries relative positions.
> The information is in topo_symbol field. I could set values of that
> field to get visually the same as what is now in GRASS. I just don't
> believe that it is better.
>
> Radim