[pgrouting-users] TSP and how to contribute

[Daniel Kastl]

>
> This is why we need a function that returns the solved dijkstra tree like:
>
> node_id, cost, parent_node_id
>
> for all nodes in the solution. You can then retrieve the
> edge(parent_node_id, node_id). cost is the accumulated cost to get to
> node_id from the start node. parent_node_id should be NULL or -1 to indicate
> the node_id is the start node.
>

Good point. I agree.
Since the returned "edge" column is more or less useless, don't you think it
would fit in the current schema?
http://www.pgrouting.org/docs/1.x/dd.html

Just replace edge_id column with parent_node_id and cost
as accumulated cost.
That said, I unfortunately have no knowledge in C/C++. If someone thinks,
this isn't difficult, please go ahead!



>
> This allows anyone to do a Dijkstra solution and get back the full results
> of that solution. One of the goal of pgRouting is to better support
> arbitrary graph analysis within the database and this would greatly help in
> this area. We need the high level wrapper functions like we already have
> because they are convenient, but we currently hide access to the underlying
> graph solutions so people can not use it to do things like Emre is trying to
> do.
>

Also agree.
Do you think this applies also to shortest path functions or are they "open"
enough?
Probably we should add some tickets.




-- 
Georepublic UG & Georepublic Japan
eMail: daniel.kastl at georepublic.de
Web: http://georepublic.de
-------------- next part --------------
An HTML attachment was scrubbed...
URL: http://lists.osgeo.org/pipermail/pgrouting-users/attachments/20110115/f2dc0b09/attachment.html