[GRASS-user] overlapping areas seem valid to v.build: why?

[Markus Metz]

On Fri, Dec 2, 2011 at 1:15 PM, Maris Nartiss <maris.gis at gmail.com> wrote:
> Hello,
> that makes sense.
>
> Might I suggest to take a look at JTS Topology Suite Technical
> Specifications fig. 6 that seems to display most common issues that
> could be eliminated with GRASS cleaning tools too.

Thanks for the suggestion! Fig. 6 shows what I tried to explain, and some more.

Markus M
>
> Maris.
>
> 2011/12/2 Markus Metz <markus.metz.giswork at googlemail.com>:
>> On Fri, Dec 2, 2011 at 12:43 PM, Maris Nartiss <maris.gis at gmail.com> wrote:
>>> Pardon my ignorance,
>>> how is possible to distinguish areas without centroids from holes?
>>> And why it doesn't make sense?
>>
>> A hole is a nothing in a something. If there is no something all
>> around the hole there is no hole. IOW, an area without a centroid is a
>> hole if it is completely surrounded by other areas with centroids,
>> otherwise that area is not a hole because it shares one or more
>> boundaries with the outer void. This does not make sense logically
>> because then there are two indistinguishable voids bordering each
>> other. These two voids can just as well be combined into one, lest
>> there is something special about one of the two voids, but that you
>> could only tell if there is a centroid in one void, rendering it
>> non-void.
>>
>> Markus M
>>>
>>> Maris,
>>> just finished converting and cleaning CAD (spaghetti) lines + points
>>> to valid areas.
>>>
>>> 2011/12/2 Markus Metz <markus.metz.giswork at googlemail.com>:
>>>>
>>>>  * for areas without centroids that are not holes. Areas without
>>>> centroids that are not holes are topologically correct but do not make
>>>> sense logically.