CGAL users discussion list

[Mattia Penati <>]

On 28 March 2011 17:23, Sebastien Loriot (GeometryFactory)
<>
 wrote:
> Mattia Penati wrote:
>>
>> On 28 March 2011 15:11, Sebastien Loriot (GeometryFactory)
>> <>
>>  wrote:
>>>
>>> Hello Mattia,
>>>
>>> Conceptually I have a problem defining a graph on a data structure which
>>> is documented as a point set on which you can do queries. I mean that
>>> there is no notion of graph on the structure, whereas there is an
>>> obvious one on triangulation.
>>
>>
>> I completely agree with you about the fact that a point set doesn't
>> define a graph, but there are some applications where this definition
>> is implicit. Currently I'm working on a problem in which I need to do
>> geometric queries and to construct the Euclidean minimum spanning tree
>> (EMSP) of a 2D point set. To define the EMST obviously you need a
>> graph, but it's the complete graph given by points, so it's quite
>> useless to describe each edge. And the algorithm to construct the EMST
>> use the fact that EMST is a subgraph of Delaunay triangulation.
>>
> My answer was related to the fact that I was not enthusiast to add
> the graph_traits specialization and document it, because as you said you
> are using the underlying Delaunay triangulation. Hopefully your code
> works with the Delaunay specialization (up to documentation problem).

Yes, as I say in the previous message it works. And certainly the your
is the best solution.

(I'm sorry if I reply to you two times, but from this message I can
not understand if you read the second part of my previous reply)

-- 
Mattia Penati