<div dir="ltr"><div>Hi Dave ,<br><br>
<span style="color:rgb(0,0,255)">Can I suggest that a 'test grid' be created for this project and be made part of pgroute package.<br>
<br>
In that way, when somebody is having a problem, they can be asked to run
 a problem basied on the 'test gird' and it can be discovered very 
quickly if there is problem with the pgroute package or something else.</span><div class=""><div id=":1k3" class="" tabindex="0"><img class="" src="https://mail.google.com/mail/u/0/images/cleardot.gif"></div></div> Can you be a bit more elaborate about this??   <br>
<br></div><div>Sorry for replying late.<br><br><br></div><div>Mukul<br></div><div><br></div><div><br></div></div><div class="gmail_extra"><br><br><div class="gmail_quote">On Tue, Apr 16, 2013 at 9:00 PM, Stephen Woodbridge <span dir="ltr"><<a href="mailto:woodbri@swoodbridge.com" target="_blank">woodbri@swoodbridge.com</a>></span> wrote:<br>
<blockquote class="gmail_quote" style="margin:0 0 0 .8ex;border-left:1px #ccc solid;padding-left:1ex"><div class="im">On 4/15/2013 6:36 AM, Mukul priya wrote:<br>
<blockquote class="gmail_quote" style="margin:0 0 0 .8ex;border-left:1px #ccc solid;padding-left:1ex">
Hi Steve ,<br>
               Just to clear things i would prefer item 2 ( partitioning<br>
graph and then routing ) over item1 ( multi point routing ) .Its kind of<br>
exciting and challenging too. what are your thoughts regarding this??<br>
</blockquote>
<br></div>
I am fine with this preference. I think it is more important that whatever you work on will be exciting and keep you interested through out the course of the project.<div class="im"><br>
<br>
<blockquote class="gmail_quote" style="margin:0 0 0 .8ex;border-left:1px #ccc solid;padding-left:1ex">
            I went through the mail archive of previous years and it was<br>
quite useful .  For now i am trying to get familiar with the development<br>
framework whenever i get time  using this link<br>
(<a href="https://github.com/pgRouting/pgrouting/wiki/Developer---Getting-Started" target="_blank">https://github.com/pgRouting/<u></u>pgrouting/wiki/Developer---<u></u>Getting-Started</a>).<br>
Once i am done with my semester by 23rd of this month i will speed up<br>
significantly.<br>
<br>
              Meanwhile feedbacks and suggestions are welcome.<br>
</blockquote>
<br></div>
As you have time learning the development environment and github are critical so you can focus on your project and not the infrastructure.<br>
<br>
You should also look over:<br>
<a href="https://github.com/pgRouting/pgrouting/wiki/2.0-Development-Guidelines-and-Standards" target="_blank">https://github.com/pgRouting/<u></u>pgrouting/wiki/2.0-<u></u>Development-Guidelines-and-<u></u>Standards</a><br>

<br>
Thanks,<br>
  -Steve<br>
<br>
<blockquote class="gmail_quote" style="margin:0 0 0 .8ex;border-left:1px #ccc solid;padding-left:1ex"><div class="im">
Thanks<br>
<br>
Mukul<br>
<br>
<br>
<br>
<br>
On Sun, Apr 14, 2013 at 7:11 PM, Mukul priya <<a href="mailto:mukul2047@gmail.com" target="_blank">mukul2047@gmail.com</a><br></div><div class="im">
<mailto:<a href="mailto:mukul2047@gmail.com" target="_blank">mukul2047@gmail.com</a>>> wrote:<br>
<br>
    Thanks  for the reply Steve , clarifies all the issues that i raised<br>
    , Proposed data structures cover  what we need , quad tree should<br>
    work too ,  I am right now looking into the last part which<br>
    describes the method of appending our graph with new cells , seems<br>
    fine and very much implementable , will post something in case some<br>
    new ideas strike me . :)<br>
<br>
<br>
      Thanks.<br>
<br>
<br>
    On Sun, Apr 14, 2013 at 10:59 AM, Stephen Woodbridge<br></div><div><div class="h5">
    <<a href="mailto:woodbri@swoodbridge.com" target="_blank">woodbri@swoodbridge.com</a> <mailto:<a href="mailto:woodbri@swoodbridge.com" target="_blank">woodbri@swoodbridge.<u></u>com</a>>> wrote:<br>
<br>
        On 4/13/2013 11:12 PM, Mukul priya wrote:<br>
<br>
            Thanks Steve for clearly mentioning the two proposals.<br>
<br>
            For Item 1 , upgrading all the algorithms will certainly<br>
            require a lot<br>
            of work and since i will be having my summer vacation i<br>
            don't have any<br>
            issue with it :).<br>
<br>
<br>
            For item 2 , i am looking into the idea that you have<br>
            proposed , It is<br>
            very much doable , there are however some concerns of mine like<br>
<br>
            - how do we decide what should be the grid size . this can<br>
            vary for<br>
            suburban area and urban area based on netwrok density.<br>
<br>
<br>
        This might be done with a quad tree approach. You start with a<br>
        square and some condition like maximum number of node. If you<br>
        exceed that number you quarter it into 4 squares and divide the<br>
        point into each of them.<br>
<br>
<br>
            - how do we classify the nodes lying on the junction of two<br>
            or more<br>
            grids . should it be assigned to all the grids??<br>
<br>
<br>
        A node can only lie in one square or the edge boundary of a<br>
        square it does not matter which one it is put in. Edges need to<br>
        be flagged if the cross a square boundary.<br>
<br>
<br>
            - how do we decide the grid that needs to be loaded in the<br>
            memory ,<br>
            connectivity with the previous grid seems legit here but i<br>
            guess we need<br>
            to discuss some corner cases too.<br>
<br>
<br>
        We could probably do something like<br>
<br>
        typedef struct pair {<br>
           int a;<br>
           int b;<br>
        } PAIR;<br>
<br>
        typedef struct edge_type {<br>
           int node_a;<br>
           int node_b;<br>
           PAIR *squares; // NULL if it does not cross square edge<br>
           float cost;<br>
           float rcost;<br>
        } EDGE;<br>
<br>
        Where PAIR can be assign the gridcell for the a and b ends.<br>
<br>
        If we number the grid cells by their quadtree numbers like:<br>
<br>
        +---+---+<br>
        | 1 | 2 |<br>
        +---+---+<br>
        | 3 | 4 |<br>
        +---+---+<br>
<br>
        So you start, with the above for your coverage area. So all<br>
        nodes would fall into cells 1-4. If you had to split cell 1<br>
        above, then those 4 new cells would be number 11, 12, 13, 14 and<br>
        the remaining unsplit cells would still be 2, 3, 4. If you had<br>
        to further split cell 14, then the new cells would be numbered<br>
        141, 142, 143, 144. So each time a cell gets subdivided, it gets<br>
        another digit added.<br>
<br>
        This is the challenge of design good algorithms, if we have<br>
        millions of edges and node, we need to be efficient memory wise<br>
        with our data structures but still be fast. In the database, you<br>
        need to think about where the data is coming from (ie: tables<br>
        using queries) and when it gets moved into memory. You can't<br>
        think just C code or database code, you have to use both.<br>
<br>
        The idea being that we want to prepare our data in tables, then<br>
        issue queries from C to get the new edges we need to append<br>
        cell(s) to our graph. So I'm thinking that we have a recursive<br>
        plpgsql procedure that splits the nodes into the appropriate<br>
        quadtree cells based on some rules. So for example we have a<br>
        vertices_tmp table that we use to assign node numbers to nodes,<br>
        we could add a cell column like this:<br>
<br>
        CREATE TABLE vertices_tmp<br>
        (<br>
           id serial NOT NULL,<br>
           cell bigint,<br>
           the_geom geometry,<br>
        );<br>
<br>
        and after we run the quadtree analysis each node is assigned a<br>
        cell number. The edge table has node_a and node_b assigned to it<br>
        also.<br>
<br>
        If we want all edges related to cell 114 then we can do a query<br>
        like:<br>
<br>
        select b.*<br>
           from vertices_tmp a, edges b<br></div></div>
          where a.cell=114 and (<a href="http://a.id" target="_blank">a.id</a> <<a href="http://a.id" target="_blank">http://a.id</a>>=b.node_a or <a href="http://a.id" target="_blank">a.id</a><br>
        <<a href="http://a.id" target="_blank">http://a.id</a>>=b.node.b);<div class="im"><br>
<br>
        Thoughts?<br>
<br>
        -Steve<br>
<br>
<br>
            Thanks.<br>
<br>
<br>
<br>
<br>
            On Sat, Apr 13, 2013 at 6:20 AM, Stephen Woodbridge<br>
            <<a href="mailto:woodbri@swoodbridge.com" target="_blank">woodbri@swoodbridge.com</a> <mailto:<a href="mailto:woodbri@swoodbridge.com" target="_blank">woodbri@swoodbridge.<u></u>com</a>><br></div><div class="im">

            <mailto:<a href="mailto:woodbri@swoodbridge." target="_blank">woodbri@swoodbridge.</a>__<u></u>com<br>
            <mailto:<a href="mailto:woodbri@swoodbridge.com" target="_blank">woodbri@swoodbridge.<u></u>com</a>>>> wrote:<br>
<br></div><div><div class="h5">
                 On 4/11/2013 3:24 PM, Mukul priya wrote:<br>
<br>
                     Hi Steve and Daniel,<br>
<br>
<br>
                               You suggested extending the present<br>
            algorithms such<br>
                     that its<br>
                     input  can take more points and not only the source and<br>
                     destination . i<br>
                     think this can be implemented and i will soon come<br>
            up with<br>
                     implementation details( kind of more technical ) .<br>
<br>
                               Can you be a liitle bit more elaborate about<br>
                     partioning data<br>
                     into spatial chunks or even suggest some readings .<br>
            I can then<br>
                     come up<br>
                     with some better ideas about implementing it.<br>
<br>
<br>
                 This is just me thinking out loud :)<br>
<br>
                 Lets say we construct a grid of one degree squares, you can<br>
                 visualize it by drawing a grid over your map.<br>
<br>
                 Then you can associate all the nodes with the grid they<br>
            fall in. We<br>
                 might also need to associate edges to the grids also<br>
            and an edge the<br>
                 crosses over a grid boundary might need to be<br>
            associated with two<br>
                 grids. This could be done with some simple relational<br>
            tables like:<br>
<br>
                 node_id|grid_id  or  edge_d|grid_id<br>
<br>
                 So my idea would be to do the routing like this:<br>
<br>
                 1. get the start node or edge<br>
                 2. build the graph based on loading the items in the<br>
            related grid<br>
                 3. mark the boundary nodes (we might want to do this<br>
            when we grid them)<br>
                 4. run the algorithm until we find the target node or<br>
            hit a boundary<br>
                 node<br>
                 5. on hitting a boundary:<br>
                    a. we check if the connected grid is loaded and<br>
            continue if it is<br>
                    b. or we extent the graph with the new grid<br>
                 6. continue with the routing<br>
<br>
                 We might want to be able to dynamically change the size<br>
            of the grid<br>
                 cell based on the number of items in it. This would<br>
            give us better<br>
                 performance when transitioning from rural areas into<br>
            urban areas<br>
                 where there is a greater density of road segments.<br>
            Think of a<br>
                 quadtree where we split it based on number of entities.<br>
<br>
<br>
                               Daniel , i took a look at the oracle link<br>
            that you<br>
                     provided but<br>
                     there was no details about how it has been<br>
            implemented , it was more<br>
                     about how one can use it. May be a bit  more search<br>
               and luck might<br>
                     take me to its implementation document :) .<br>
<br>
<br>
                 Right, it is useful if you look at the documentation<br>
            and ask why did<br>
                 they do it that way and what does it imply about how it<br>
            works behind<br>
                 the scenes.<br>
<br>
<br>
                                The other thing that you mentioned was<br>
            about contraction<br>
                     Hierarchy . Still the nodes have to be ordered<br>
            based on some<br>
                     criteria or<br>
                     according to their importance . We can  use natural<br>
            hierarchy<br>
                     present in<br>
                     the network for doing that .<br>
<br>
<br>
                 This is not related to what you proposed. It is an<br>
            algorithm that<br>
                 does a lot of precompuation, that is LOTS in capitals,<br>
            but it can<br>
                 get results in milliseconds for cross country routes.<br>
<br>
<br>
                                 i will be really grateful if anyone can<br>
            correct me<br>
                     in case if<br>
                     my thought process in not on the right lane and<br>
            sorry for the<br>
                     late reply<br>
                     as my academic session  is in progress too<br>
            .Meanwhile  i am<br>
                     trying to<br>
                     get fluent with git ,cmake and other tools.<br>
<br>
<br>
                 So read over our wiki:<br>
<br></div></div>
            <a href="https://github.com/pgRouting/____pgrouting/wiki" target="_blank">https://github.com/pgRouting/_<u></u>___pgrouting/wiki</a><br>
            <<a href="https://github.com/pgRouting/__pgrouting/wiki" target="_blank">https://github.com/pgRouting/<u></u>__pgrouting/wiki</a>><div class="im"><br>
<br>
                 <<a href="https://github.com/pgRouting/__pgrouting/wiki" target="_blank">https://github.com/pgRouting/<u></u>__pgrouting/wiki</a><br>
            <<a href="https://github.com/pgRouting/pgrouting/wiki" target="_blank">https://github.com/pgRouting/<u></u>pgrouting/wiki</a>>><br>
<br>
                 The way I see it at the moment there are two unrelated<br>
            proposals on<br>
                 the table (I'm leaving out the contraction hierarchy):<br>
<br>
                 1. multi point routing<br>
                 2. partition JIT graph building while routing<br>
<br>
                 Item 1 is fairly trivial technically, I think, but if<br>
            you were to<br>
                 upgrade all the algorithms to do this it would be a lot<br>
            of work and<br>
                 a useful contribution to pgrouting.<br>
<br>
                 Item 2 is more of a design and code a new algorithm and<br>
            you would<br>
                 probably want to focus on using astar or trsp algorithm<br>
            to do this<br>
                 with. This one is more technically challenging and has<br>
            more unknowns<br>
                 in it but I think it should be doable.<br>
<br>
                 If you are interested in reading about contraction<br>
            hierarchies:<br></div>
            <a href="https://www.google.com/#q=____contraction" target="_blank">https://www.google.com/#q=____<u></u>contraction</a><br>
            <<a href="https://www.google.com/#q=__contraction" target="_blank">https://www.google.com/#q=__<u></u>contraction</a>><div class="im"><br>
<br>
                 <<a href="https://www.google.com/#q=__contraction" target="_blank">https://www.google.com/#q=__<u></u>contraction</a><br>
            <<a href="https://www.google.com/#q=contraction" target="_blank">https://www.google.com/#q=<u></u>contraction</a>>> hierarchies<br>
<br>
                 Thanks,<br>
                    -Steve<br>
<br>
                     Thanks .<br>
<br>
                     Mukul<br>
<br>
<br>
<br>
<br>
                     On Thu, Apr 11, 2013 at 6:47 PM, Stephen Woodbridge<br>
                     <<a href="mailto:woodbri@swoodbridge.com" target="_blank">woodbri@swoodbridge.com</a><br>
            <mailto:<a href="mailto:woodbri@swoodbridge.com" target="_blank">woodbri@swoodbridge.<u></u>com</a>><br>
            <mailto:<a href="mailto:woodbri@swoodbridge." target="_blank">woodbri@swoodbridge.</a>__<u></u>com<br>
            <mailto:<a href="mailto:woodbri@swoodbridge.com" target="_blank">woodbri@swoodbridge.<u></u>com</a>>><br></div><div><div class="h5">
                     <mailto:<a href="mailto:woodbri@swoodbridge" target="_blank">woodbri@swoodbridge</a>.<br>
            <mailto:<a href="mailto:woodbri@swoodbridge" target="_blank">woodbri@swoodbridge</a>.>_<u></u>___com<br>
                     <mailto:<a href="mailto:woodbri@swoodbridge." target="_blank">woodbri@swoodbridge.</a>__<u></u>com<br>
            <mailto:<a href="mailto:woodbri@swoodbridge.com" target="_blank">woodbri@swoodbridge.<u></u>com</a>>>>> wrote:<br>
<br>
                          With pgRouting, we do most things dynamically,<br>
            here is the<br>
                     basic flow:<br>
<br>
                          1. Given a collection of input, points, nodes,<br>
            or edges<br>
                          map these to nodes or edges depending on<br>
            algorithm.<br>
<br>
                          2. Select the edges we need to build the graph<br>
<br>
                          3. build the graph and solve it<br>
<br>
                          4. return the results<br>
<br>
                          All our algorithms today only take start and<br>
            end points.<br>
                     They could<br>
                          be extended to take points. Each "point" (I<br>
            use "point" as<br>
                     a generic<br>
                          term because it might be a lat/lon, node id,<br>
            edge id and<br>
                     offset,<br>
                          etc) would need to be mapped to the<br>
            appropriate input need<br>
                     for any<br>
                          given algorithm.<br>
<br>
                          So for node based algorithms like Dijkstra,and<br>
            astar it<br>
                     would get<br>
                          resolved to a node. For TRSP it would get<br>
            mapped to the<br>
                     nearest edge<br>
                          and offset along that edge. Postgis has lots<br>
            of handy tools for<br>
                          doing this.<br>
<br>
                          -Steve<br>
<br>
<br>
                          On 4/10/2013 10:50 PM, Mukul priya wrote:<br>
<br>
                              Thanks for the reply steve . i have<br>
            already cloned the<br>
                     github<br>
                              repository<br>
                              and looking into various aspects of it .<br>
<br>
                              For IRNN querry implementation i think it<br>
            is a good<br>
                     idea to sub<br>
                              divide<br>
                              the whole route and generate n+1 routes<br>
            separately ,<br>
                     say from S<br>
                              to F1 ,<br>
                              F1-F2 ,..... F(n-1)-Fn , Fn to D . i<br>
            wanted to know if<br>
                     we have<br>
                              that kind<br>
                              of a data where each and every facility is<br>
            mentioned on<br>
                     the map as a<br>
                              point (node ) . even if it is not directly<br>
            connected to<br>
                     the road<br>
                              network<br>
                              we can basically treat it a pseudo node<br>
            and then call<br>
                     the router<br>
                              . The<br>
                              other thing about optimization yes we can<br>
            do that using<br>
                     spatial<br>
                              range<br>
                              querries suppose there are several<br>
            instances of the same<br>
                              facility that a<br>
                              user wants to access then we can use<br>
            spatial range<br>
                     querries to<br>
                              locate<br>
                              that facility which is the nearest.<br>
<br>
<br>
                              On Thu, Apr 11, 2013 at 1:42 AM, Stephen<br>
            Woodbridge<br>
                              <<a href="mailto:woodbri@swoodbridge.com" target="_blank">woodbri@swoodbridge.com</a><br>
            <mailto:<a href="mailto:woodbri@swoodbridge.com" target="_blank">woodbri@swoodbridge.<u></u>com</a>><br>
                     <mailto:<a href="mailto:woodbri@swoodbridge." target="_blank">woodbri@swoodbridge.</a>__<u></u>com<br>
            <mailto:<a href="mailto:woodbri@swoodbridge.com" target="_blank">woodbri@swoodbridge.<u></u>com</a>>><br>
                     <mailto:<a href="mailto:woodbri@swoodbridge" target="_blank">woodbri@swoodbridge</a>.<br>
            <mailto:<a href="mailto:woodbri@swoodbridge" target="_blank">woodbri@swoodbridge</a>.>_<u></u>___com<br>
            <mailto:<a href="mailto:woodbri@swoodbridge." target="_blank">woodbri@swoodbridge.</a>__<u></u>com<br>
            <mailto:<a href="mailto:woodbri@swoodbridge.com" target="_blank">woodbri@swoodbridge.<u></u>com</a>>>><br></div></div>
                              <mailto:<a href="mailto:woodbri@swoodbridge" target="_blank">woodbri@swoodbridge</a><br>
            <mailto:<a href="mailto:woodbri@swoodbridge" target="_blank">woodbri@swoodbridge</a>>.<br>
                     <mailto:<a href="mailto:woodbri@swoodbridge" target="_blank">woodbri@swoodbridge</a><br>
            <mailto:<a href="mailto:woodbri@swoodbridge" target="_blank">woodbri@swoodbridge</a>>.><u></u>______com<div><div class="h5"><br>
<br>
<br>
                              <mailto:<a href="mailto:woodbri@swoodbridge" target="_blank">woodbri@swoodbridge</a>.<br>
            <mailto:<a href="mailto:woodbri@swoodbridge" target="_blank">woodbri@swoodbridge</a>.>_<u></u>___com<br>
                     <mailto:<a href="mailto:woodbri@swoodbridge." target="_blank">woodbri@swoodbridge.</a>__<u></u>com<br>
            <mailto:<a href="mailto:woodbri@swoodbridge.com" target="_blank">woodbri@swoodbridge.<u></u>com</a>>>>>> wrote:<br>
<br>
                                   On 4/10/2013 3:23 PM, Mukul priya wrote:<br>
<br>
<br>
                                       Hi ,<br>
<br>
<br>
                                   Hi Mukul,<br>
<br>
                                   Thank you for your interest in pgRouting.<br>
<br>
                                                      I am a  B.tech<br>
            fourth year<br>
                     student at<br>
                              IIIT-Hyderabad<br>
                                       pursuing a degree in computer<br>
            science and<br>
                     engineering<br>
                              and i will<br>
                                       be soon<br>
                                       pursuing a Masters Degree in the<br>
            field of Spatial<br>
                              Informatics<br>
                                       and the<br>
                                       research topic that i have been<br>
            working on is *"In<br>
                              route nearest<br>
                                       neighbour querries".*<br>
<br>
<br>
                                                      Last year i worked<br>
            on a project<br>
                     that was<br>
                              funded by<br>
                                       Honeywell technology solutions<br>
            and it gave me<br>
                     a lot of<br>
                              insight about<br>
                                       open source programming and<br>
            industrial work<br>
                     culture.<br>
<br>
                                       I was introduced to pgrouting by<br>
            *Prof. Venkatesh<br>
                              Raghavan* who<br>
                                       visited<br>
<br>
                                       our college last summer. i have<br>
            also used<br>
                     pgrouting for<br>
                                       implementing one<br>
                                       of my Honors project.<br>
<br>
                                       i have gone through the updated<br>
            ideas page and<br>
                     i am<br>
                              listing out<br>
                                       a topic<br>
                                       that i feel i can contribute to.<br>
<br>
                                       *Idea *<br>
<br>
                                       Network Partitioning<br>
<br>
                                       A very simple method using which<br>
            it can be<br>
                     done is :<br>
<br>
                                          * *Existence of a natural<br>
            Hierarchy*<br>
<br>
<br>
                                                  Generally road<br>
            networks are<br>
                     organized such<br>
                              that there<br>
                                       is some<br>
                                       natural hierarchy for example if<br>
            we look at<br>
                     the road<br>
                              network of<br>
                                       USA we<br>
                                       observe that there are national<br>
            highways which<br>
                     connect<br>
                              multiple<br>
                                       large<br>
                                       regions , inter state roads<br>
            connect places<br>
                     within these<br>
                              regions<br>
                                       , multi<br>
                                       lane roads connect city areas and<br>
            then there<br>
                     are small<br>
                              roads to<br>
                                       connect<br>
                                       individual houses.<br>
<br>
                                                    so what we can do is<br>
            first rank these<br>
                              classes that<br>
                                       constitute the road network and<br>
            then use the<br>
                     highest<br>
                              level roads to<br>
                                       divide the road network into<br>
            large regions<br>
                     enclosed by<br>
                              these<br>
                                       roads. each<br>
                                       of the divided regions can<br>
            further be divided<br>
                     again<br>
                              using next lower<br>
                                       level road.<br>
<br>
                                                    so suppose we have a<br>
            road network<br>
                     which n<br>
                              classes of<br>
                                       different roads then we can<br>
            create a tree of<br>
                     depth n-1<br>
                              where the<br>
                                       root of<br>
                                       the tree will represent the<br>
            entire road<br>
                     network and<br>
                              children of<br>
                                       the the<br>
                                       root node will represent the area<br>
            formed by<br>
                              partitioning the<br>
                                       root using<br>
                                       the level 1 ( highest ) edges and<br>
            so on . the<br>
                     nodes<br>
                              will basically<br>
                                       represent a smaller part of the<br>
            road network.<br>
<br>
                                                     The idea seems to<br>
            be very naive<br>
                     right now<br>
                              but if<br>
                                       anyone can<br>
                                       give some feedback whether it is<br>
            achievable or<br>
                     not or<br>
                              may be suggest<br>
                                       some modifications.<br>
<br>
<br>
                                   Yes this is the basics of how this<br>
            could work.<br>
                     Because we<br>
                              build our<br>
                                   graphs dynamically for each route<br>
            request, we can do<br>
                              something like<br>
                                   this today. Typically you have to<br>
            feed the route<br>
                     request<br>
                              and SQL<br>
                                   query that provides the edges needed<br>
            to build the<br>
                     graph and<br>
                              this can<br>
                                   be simply the bounding box of the<br>
            start and end<br>
                     point of<br>
                              the route<br>
                                   expanded slightly to allow the route<br>
            move outside that<br>
                              bounds by a<br>
                                   little if needed. A case in point are<br>
            start and<br>
                     end points<br>
                              that form<br>
                                   a vertical of horizontal line.<br>
<br>
                                   So for the natural hierarchy, you can<br>
            for a SQL<br>
                     query like:<br>
<br>
                                   select * from edges where<br>
            st_dwithin(the_geom,<br>
                     start_pnt,<br>
                              radius)<br>
                                   union<br>
                                   select * from edges where<br>
            st_dwithin(the_geom,<br>
                     end_pnt, radius)<br>
                                   union<br>
                                   select * from edges<br>
                                      where<br></div></div>
            st_expand(st_makeline(start___<u></u>______pnt,<div><div class="h5"><br>
<br>
                     end_pnt), pct)<br>
<br>
<br>
                                        and road_class < 4;<br>
<br>
                                   So this gets all edges regardless of<br>
            class at the<br>
                     start and<br>
                              end and<br>
                                   then gets all the major roads and<br>
            highways between the<br>
                              start and end<br>
                                   points. We can dynamically select the<br>
            edges that<br>
                     we want<br>
                              when we<br>
                                   build the graph.<br>
<br>
                                   Regardless of how you implement the<br>
            routing, the<br>
                     problem is all<br>
                                   about the data. If you have a road<br>
            segment the is<br>
                              misqualified, you<br>
                                   might end up with a network that is<br>
            broken between<br>
                     start<br>
                              and end.<br>
                                   This can alsoo happen if ramps are<br>
            not coded<br>
                     correctly.<br>
<br>
                                   One of the challenges we have today<br>
            is that we<br>
                     have to be<br>
                              able to<br>
                                   construct the whole graph in memory<br>
            before we can<br>
                     start<br>
                              routing.<br>
                                   This is ok for small areas but it is<br>
            a problem if<br>
                     you want to<br>
                                   generate a route between say Miami,<br>
            Florida and<br>
                     Seattle,<br>
                              Washington.<br>
                                   An interesting problem would be the<br>
            ability to<br>
                     partition<br>
                              the data in<br>
                                   spatial chucks and only load them as<br>
            the solver<br>
                     needed them.<br>
<br>
                                   If you think about your edges sorted<br>
            into say 1<br>
                     degree grid<br>
                              partitions,<br>
                                   then you load the partition for the<br>
            start point<br>
                     and start<br>
                              routing<br>
                                   using A* search, when you frontier<br>
            get to an edge<br>
                     of the<br>
                              grid you<br>
                                   are in, then you load the adjacent<br>
            grid and<br>
                     continue, if<br>
                              you bump<br>
                                   into another grid boundary that is<br>
            not loaded yet, you<br>
                              load, if it<br>
                                   is already loaded you continue.<br>
            Anyway food for<br>
                     thought! :)<br>
<br>
<br>
                                                      In route nearest<br>
            neighbour<br>
                     querries(<br>
                              IRNN) which<br>
                                       handle<br>
                                       querries like computation of<br>
            shortest path ,<br>
                     provided<br>
                              that the user<br>
                                       wants to visit facilities F1 , F2<br>
            ,.....FN<br>
                     while he/she<br>
                              drives<br>
                                       or walks<br>
                                       from source to destination. Network<br>
                     partitioning can<br>
                              optimize these<br>
                                       computations  too as the search<br>
            space reduces<br>
                              significantly once<br>
                                       we have<br>
                                       the partitions. Handling such<br>
            querries have<br>
                     not been<br>
                              implemented<br>
                                       yet. It<br>
                                       will be very helpful if we can<br>
            have some<br>
                     discussion<br>
                              about whether<br>
                                       implementing it is feasible or not.<br>
<br>
<br>
                                   What is we just added via support to<br>
            routing? Then<br>
                     we could do<br>
                                   something like say generate a route:<br>
            Start, F1,<br>
                     F2, ... Fn, End<br>
                                   This would allow us to build a graph<br>
            one time and<br>
                     then generate<br>
                                   multiple sub-routes with in the<br>
            graph. Today if<br>
                     you want to<br>
                              do that<br>
                                   you have to generate n+1 routes and<br>
            build the<br>
                     graph n+1<br>
                              times. We<br>
                                   could also do some preliminary<br>
            optimization of the via<br>
                              points based<br>
                                   on Euclidean distance using something<br>
            like TSP before<br>
                              calling the<br>
                                   router.<br>
<br>
<br>
<br>
                                                      It would be great<br>
            if someone<br>
                     could give<br>
                              a general<br>
                                       idea<br>
                                       how  to go about learning more<br>
            about the areas<br>
                              mentioned  with<br>
                                       respect<br>
                                       to the organization's<br>
            projects.Specially<br>
                     suggest those<br>
                              ideas<br>
                                       which the<br>
                                       developers think are achievable<br>
            for now . I<br>
                     will also<br>
                              be grateful if<br>
                                       somebody can guide me regarding<br>
            the development<br>
                              framework of<br>
                                       pgrouting<br>
                                       so that i get familiar with the whole<br>
                     framework in the<br>
                              coming days.<br>
<br>
<br>
                                   I would clone the github repository<br>
            and look at branch<br>
                              sew-devel-2_0<br>
                                   this is our new tree structure and it<br>
            has code,<br>
                     doc, and<br>
                              test all<br>
                                   organized in a nice way that makes it<br>
            easy to multiple<br>
                              contributors<br>
                                   work with the tree.<br>
<br>
                                   Ask questions, There is a tutorial<br>
            floating around<br>
                     and lots of<br>
                                   people that are will to help.<br>
<br>
                                   -Steve<br>
<br>
                                       Thank you .<br>
<br>
                                       Mukul Priya<br>
                                       Lab for spatial Informatics<br>
                                       IIIT-Hyderabad<br>
<br>
<br>
<br>
<br></div></div>
            ______________________________<u></u>_________________________<div class="im"><br>
<br>
<br>
                                       pgrouting-dev mailing list<br>
            <a href="mailto:pgrouting-dev@lists.osgeo.org" target="_blank">pgrouting-dev@lists.osgeo.org</a><br>
            <mailto:<a href="mailto:pgrouting-dev@lists.osgeo.org" target="_blank">pgrouting-dev@lists.<u></u>osgeo.org</a>><br>
            <mailto:<a href="mailto:pgrouting-dev@lists." target="_blank">pgrouting-dev@lists.</a>__<a href="http://osgeo.org" target="_blank"><u></u>osgeo.org</a><br>
            <mailto:<a href="mailto:pgrouting-dev@lists.osgeo.org" target="_blank">pgrouting-dev@lists.<u></u>osgeo.org</a>>><br>
                              <mailto:<a href="mailto:pgrouting-dev@lists" target="_blank">pgrouting-dev@lists</a>.<br>
            <mailto:<a href="mailto:pgrouting-dev@lists" target="_blank">pgrouting-dev@lists</a>.>_<u></u>___<a href="http://osgeo.org" target="_blank">osgeo.org</a> <<a href="http://osgeo.org" target="_blank">http://osgeo.org</a>><br>

                     <mailto:<a href="mailto:pgrouting-dev@lists." target="_blank">pgrouting-dev@lists.</a>__<a href="http://osgeo.org" target="_blank"><u></u>osgeo.org</a><br>
            <mailto:<a href="mailto:pgrouting-dev@lists.osgeo.org" target="_blank">pgrouting-dev@lists.<u></u>osgeo.org</a>>>><br></div>
                              <mailto:<a href="mailto:pgrouting-dev@lists" target="_blank">pgrouting-dev@lists</a><br>
            <mailto:<a href="mailto:pgrouting-dev@lists" target="_blank">pgrouting-dev@lists</a>>.<br>
                     <mailto:<a href="mailto:pgrouting-dev@lists" target="_blank">pgrouting-dev@lists</a><br>
            <mailto:<a href="mailto:pgrouting-dev@lists" target="_blank">pgrouting-dev@lists</a>>.><u></u>______<a href="http://osgeo.org" target="_blank">osgeo.org</a><br>
            <<a href="http://osgeo.org" target="_blank">http://osgeo.org</a>> <<a href="http://osgeo.org" target="_blank">http://osgeo.org</a>><div class="im"><br>
                              <mailto:<a href="mailto:pgrouting-dev@lists" target="_blank">pgrouting-dev@lists</a>.<br>
            <mailto:<a href="mailto:pgrouting-dev@lists" target="_blank">pgrouting-dev@lists</a>.>_<u></u>___<a href="http://osgeo.org" target="_blank">osgeo.org</a> <<a href="http://osgeo.org" target="_blank">http://osgeo.org</a>><br>

                     <mailto:<a href="mailto:pgrouting-dev@lists." target="_blank">pgrouting-dev@lists.</a>__<a href="http://osgeo.org" target="_blank"><u></u>osgeo.org</a><br>
            <mailto:<a href="mailto:pgrouting-dev@lists.osgeo.org" target="_blank">pgrouting-dev@lists.<u></u>osgeo.org</a>>>>><br></div>
            <a href="http://lists.osgeo.org/________mailman/listinfo/pgrouting-dev" target="_blank">http://lists.osgeo.org/_______<u></u>_mailman/listinfo/pgrouting-<u></u>dev</a> <<a href="http://lists.osgeo.org/______mailman/listinfo/pgrouting-dev" target="_blank">http://lists.osgeo.org/______<u></u>mailman/listinfo/pgrouting-dev</a><u></u>><br>

<br>
            <<a href="http://lists.osgeo.org/______mailman/listinfo/pgrouting-dev" target="_blank">http://lists.osgeo.org/______<u></u>mailman/listinfo/pgrouting-dev</a><br>
            <<a href="http://lists.osgeo.org/____mailman/listinfo/pgrouting-dev" target="_blank">http://lists.osgeo.org/____<u></u>mailman/listinfo/pgrouting-dev</a><u></u>>__><div class="im"><br>
<br>
<br>
            <<a href="http://lists.osgeo.org/______mailman/listinfo/pgrouting-dev" target="_blank">http://lists.osgeo.org/______<u></u>mailman/listinfo/pgrouting-dev</a><br>
            <<a href="http://lists.osgeo.org/____mailman/listinfo/pgrouting-dev" target="_blank">http://lists.osgeo.org/____<u></u>mailman/listinfo/pgrouting-dev</a><u></u>><br>
<br>
            <<a href="http://lists.osgeo.org/____mailman/listinfo/pgrouting-dev" target="_blank">http://lists.osgeo.org/____<u></u>mailman/listinfo/pgrouting-dev</a><br></div>
            <<a href="http://lists.osgeo.org/__mailman/listinfo/pgrouting-dev" target="_blank">http://lists.osgeo.org/__<u></u>mailman/listinfo/pgrouting-dev</a><u></u>>__>__><div class="im"><br>
<br>
<br>
<br>
<br>
            <<a href="http://lists.osgeo.org/______mailman/listinfo/pgrouting-dev" target="_blank">http://lists.osgeo.org/______<u></u>mailman/listinfo/pgrouting-dev</a><br>
            <<a href="http://lists.osgeo.org/____mailman/listinfo/pgrouting-dev" target="_blank">http://lists.osgeo.org/____<u></u>mailman/listinfo/pgrouting-dev</a><u></u>><br>
<br>
            <<a href="http://lists.osgeo.org/____mailman/listinfo/pgrouting-dev" target="_blank">http://lists.osgeo.org/____<u></u>mailman/listinfo/pgrouting-dev</a><br>
            <<a href="http://lists.osgeo.org/__mailman/listinfo/pgrouting-dev" target="_blank">http://lists.osgeo.org/__<u></u>mailman/listinfo/pgrouting-dev</a><u></u>>__><br>
<br>
<br>
            <<a href="http://lists.osgeo.org/____mailman/listinfo/pgrouting-dev" target="_blank">http://lists.osgeo.org/____<u></u>mailman/listinfo/pgrouting-dev</a><br>
            <<a href="http://lists.osgeo.org/__mailman/listinfo/pgrouting-dev" target="_blank">http://lists.osgeo.org/__<u></u>mailman/listinfo/pgrouting-dev</a><u></u>><br>
<br>
            <<a href="http://lists.osgeo.org/__mailman/listinfo/pgrouting-dev" target="_blank">http://lists.osgeo.org/__<u></u>mailman/listinfo/pgrouting-dev</a><br>
            <<a href="http://lists.osgeo.org/mailman/listinfo/pgrouting-dev" target="_blank">http://lists.osgeo.org/<u></u>mailman/listinfo/pgrouting-dev</a><u></u>>__>__>__><br>
<br>
<br>
<br></div>
            ______________________________<u></u>_________________________<div class="im"><br>
<br>
<br>
                                   pgrouting-dev mailing list<br>
            <a href="mailto:pgrouting-dev@lists.osgeo.org" target="_blank">pgrouting-dev@lists.osgeo.org</a><br>
            <mailto:<a href="mailto:pgrouting-dev@lists.osgeo.org" target="_blank">pgrouting-dev@lists.<u></u>osgeo.org</a>><br>
            <mailto:<a href="mailto:pgrouting-dev@lists." target="_blank">pgrouting-dev@lists.</a>__<a href="http://osgeo.org" target="_blank"><u></u>osgeo.org</a><br>
            <mailto:<a href="mailto:pgrouting-dev@lists.osgeo.org" target="_blank">pgrouting-dev@lists.<u></u>osgeo.org</a>>><br>
                              <mailto:<a href="mailto:pgrouting-dev@lists" target="_blank">pgrouting-dev@lists</a>.<br>
            <mailto:<a href="mailto:pgrouting-dev@lists" target="_blank">pgrouting-dev@lists</a>.>_<u></u>___<a href="http://osgeo.org" target="_blank">osgeo.org</a> <<a href="http://osgeo.org" target="_blank">http://osgeo.org</a>><br>

                     <mailto:<a href="mailto:pgrouting-dev@lists." target="_blank">pgrouting-dev@lists.</a>__<a href="http://osgeo.org" target="_blank"><u></u>osgeo.org</a><br>
            <mailto:<a href="mailto:pgrouting-dev@lists.osgeo.org" target="_blank">pgrouting-dev@lists.<u></u>osgeo.org</a>>>><br></div>
                              <mailto:<a href="mailto:pgrouting-dev@lists" target="_blank">pgrouting-dev@lists</a><br>
            <mailto:<a href="mailto:pgrouting-dev@lists" target="_blank">pgrouting-dev@lists</a>>.<br>
                     <mailto:<a href="mailto:pgrouting-dev@lists" target="_blank">pgrouting-dev@lists</a><br>
            <mailto:<a href="mailto:pgrouting-dev@lists" target="_blank">pgrouting-dev@lists</a>>.><u></u>______<a href="http://osgeo.org" target="_blank">osgeo.org</a><br>
            <<a href="http://osgeo.org" target="_blank">http://osgeo.org</a>> <<a href="http://osgeo.org" target="_blank">http://osgeo.org</a>><div class="im"><br>
                              <mailto:<a href="mailto:pgrouting-dev@lists" target="_blank">pgrouting-dev@lists</a>.<br>
            <mailto:<a href="mailto:pgrouting-dev@lists" target="_blank">pgrouting-dev@lists</a>.>_<u></u>___<a href="http://osgeo.org" target="_blank">osgeo.org</a> <<a href="http://osgeo.org" target="_blank">http://osgeo.org</a>><br>

                     <mailto:<a href="mailto:pgrouting-dev@lists." target="_blank">pgrouting-dev@lists.</a>__<a href="http://osgeo.org" target="_blank"><u></u>osgeo.org</a><br>
            <mailto:<a href="mailto:pgrouting-dev@lists.osgeo.org" target="_blank">pgrouting-dev@lists.<u></u>osgeo.org</a>>>>><br></div>
            <a href="http://lists.osgeo.org/________mailman/listinfo/pgrouting-dev" target="_blank">http://lists.osgeo.org/_______<u></u>_mailman/listinfo/pgrouting-<u></u>dev</a> <<a href="http://lists.osgeo.org/______mailman/listinfo/pgrouting-dev" target="_blank">http://lists.osgeo.org/______<u></u>mailman/listinfo/pgrouting-dev</a><u></u>><br>

<br>
            <<a href="http://lists.osgeo.org/______mailman/listinfo/pgrouting-dev" target="_blank">http://lists.osgeo.org/______<u></u>mailman/listinfo/pgrouting-dev</a><br>
            <<a href="http://lists.osgeo.org/____mailman/listinfo/pgrouting-dev" target="_blank">http://lists.osgeo.org/____<u></u>mailman/listinfo/pgrouting-dev</a><u></u>>__><div class="im"><br>
<br>
<br>
            <<a href="http://lists.osgeo.org/______mailman/listinfo/pgrouting-dev" target="_blank">http://lists.osgeo.org/______<u></u>mailman/listinfo/pgrouting-dev</a><br>
            <<a href="http://lists.osgeo.org/____mailman/listinfo/pgrouting-dev" target="_blank">http://lists.osgeo.org/____<u></u>mailman/listinfo/pgrouting-dev</a><u></u>><br>
<br>
            <<a href="http://lists.osgeo.org/____mailman/listinfo/pgrouting-dev" target="_blank">http://lists.osgeo.org/____<u></u>mailman/listinfo/pgrouting-dev</a><br></div>
            <<a href="http://lists.osgeo.org/__mailman/listinfo/pgrouting-dev" target="_blank">http://lists.osgeo.org/__<u></u>mailman/listinfo/pgrouting-dev</a><u></u>>__>__><div class="im"><br>
<br>
<br>
<br>
<br>
            <<a href="http://lists.osgeo.org/______mailman/listinfo/pgrouting-dev" target="_blank">http://lists.osgeo.org/______<u></u>mailman/listinfo/pgrouting-dev</a><br>
            <<a href="http://lists.osgeo.org/____mailman/listinfo/pgrouting-dev" target="_blank">http://lists.osgeo.org/____<u></u>mailman/listinfo/pgrouting-dev</a><u></u>><br>
<br>
            <<a href="http://lists.osgeo.org/____mailman/listinfo/pgrouting-dev" target="_blank">http://lists.osgeo.org/____<u></u>mailman/listinfo/pgrouting-dev</a><br>
            <<a href="http://lists.osgeo.org/__mailman/listinfo/pgrouting-dev" target="_blank">http://lists.osgeo.org/__<u></u>mailman/listinfo/pgrouting-dev</a><u></u>>__><br>
<br>
<br>
            <<a href="http://lists.osgeo.org/____mailman/listinfo/pgrouting-dev" target="_blank">http://lists.osgeo.org/____<u></u>mailman/listinfo/pgrouting-dev</a><br>
            <<a href="http://lists.osgeo.org/__mailman/listinfo/pgrouting-dev" target="_blank">http://lists.osgeo.org/__<u></u>mailman/listinfo/pgrouting-dev</a><u></u>><br>
<br>
            <<a href="http://lists.osgeo.org/__mailman/listinfo/pgrouting-dev" target="_blank">http://lists.osgeo.org/__<u></u>mailman/listinfo/pgrouting-dev</a><br></div>
            <<a href="http://lists.osgeo.org/mailman/listinfo/pgrouting-dev" target="_blank">http://lists.osgeo.org/<u></u>mailman/listinfo/pgrouting-dev</a><u></u>>__>__>__><div><div class="h5"><br>
<br>
<br>
<br>
<br>
<br>
<br>
<br>
              ______________________________<u></u>_______________________<br>
                              pgrouting-dev mailing list<br>
            <a href="mailto:pgrouting-dev@lists.osgeo.org" target="_blank">pgrouting-dev@lists.osgeo.org</a><br>
            <mailto:<a href="mailto:pgrouting-dev@lists.osgeo.org" target="_blank">pgrouting-dev@lists.<u></u>osgeo.org</a>><br>
                     <mailto:<a href="mailto:pgrouting-dev@lists." target="_blank">pgrouting-dev@lists.</a>__<a href="http://osgeo.org" target="_blank"><u></u>osgeo.org</a><br>
            <mailto:<a href="mailto:pgrouting-dev@lists.osgeo.org" target="_blank">pgrouting-dev@lists.<u></u>osgeo.org</a>>><br>
                     <mailto:<a href="mailto:pgrouting-dev@lists" target="_blank">pgrouting-dev@lists</a>.<br>
            <mailto:<a href="mailto:pgrouting-dev@lists" target="_blank">pgrouting-dev@lists</a>.>_<u></u>___<a href="http://osgeo.org" target="_blank">osgeo.org</a> <<a href="http://osgeo.org" target="_blank">http://osgeo.org</a>><br>

                     <mailto:<a href="mailto:pgrouting-dev@lists." target="_blank">pgrouting-dev@lists.</a>__<a href="http://osgeo.org" target="_blank"><u></u>osgeo.org</a><br>
            <mailto:<a href="mailto:pgrouting-dev@lists.osgeo.org" target="_blank">pgrouting-dev@lists.<u></u>osgeo.org</a>>>><br>
            <a href="http://lists.osgeo.org/______mailman/listinfo/pgrouting-dev" target="_blank">http://lists.osgeo.org/______<u></u>mailman/listinfo/pgrouting-dev</a><br>
            <<a href="http://lists.osgeo.org/____mailman/listinfo/pgrouting-dev" target="_blank">http://lists.osgeo.org/____<u></u>mailman/listinfo/pgrouting-dev</a><u></u>><br>
<br>
            <<a href="http://lists.osgeo.org/____mailman/listinfo/pgrouting-dev" target="_blank">http://lists.osgeo.org/____<u></u>mailman/listinfo/pgrouting-dev</a><br>
            <<a href="http://lists.osgeo.org/__mailman/listinfo/pgrouting-dev" target="_blank">http://lists.osgeo.org/__<u></u>mailman/listinfo/pgrouting-dev</a><u></u>>__><br>
<br>
<br>
            <<a href="http://lists.osgeo.org/____mailman/listinfo/pgrouting-dev" target="_blank">http://lists.osgeo.org/____<u></u>mailman/listinfo/pgrouting-dev</a><br>
            <<a href="http://lists.osgeo.org/__mailman/listinfo/pgrouting-dev" target="_blank">http://lists.osgeo.org/__<u></u>mailman/listinfo/pgrouting-dev</a><u></u>><br>
<br>
            <<a href="http://lists.osgeo.org/__mailman/listinfo/pgrouting-dev" target="_blank">http://lists.osgeo.org/__<u></u>mailman/listinfo/pgrouting-dev</a><br>
            <<a href="http://lists.osgeo.org/mailman/listinfo/pgrouting-dev" target="_blank">http://lists.osgeo.org/<u></u>mailman/listinfo/pgrouting-dev</a><u></u>>__>__><br>
<br>
<br>
<br>
              ______________________________<u></u>_______________________<br>
                          pgrouting-dev mailing list<br>
            <a href="mailto:pgrouting-dev@lists.osgeo.org" target="_blank">pgrouting-dev@lists.osgeo.org</a><br>
            <mailto:<a href="mailto:pgrouting-dev@lists.osgeo.org" target="_blank">pgrouting-dev@lists.<u></u>osgeo.org</a>><br>
                     <mailto:<a href="mailto:pgrouting-dev@lists." target="_blank">pgrouting-dev@lists.</a>__<a href="http://osgeo.org" target="_blank"><u></u>osgeo.org</a><br>
            <mailto:<a href="mailto:pgrouting-dev@lists.osgeo.org" target="_blank">pgrouting-dev@lists.<u></u>osgeo.org</a>>><br>
                     <mailto:<a href="mailto:pgrouting-dev@lists" target="_blank">pgrouting-dev@lists</a>.<br>
            <mailto:<a href="mailto:pgrouting-dev@lists" target="_blank">pgrouting-dev@lists</a>.>_<u></u>___<a href="http://osgeo.org" target="_blank">osgeo.org</a> <<a href="http://osgeo.org" target="_blank">http://osgeo.org</a>><br>

                     <mailto:<a href="mailto:pgrouting-dev@lists." target="_blank">pgrouting-dev@lists.</a>__<a href="http://osgeo.org" target="_blank"><u></u>osgeo.org</a><br>
            <mailto:<a href="mailto:pgrouting-dev@lists.osgeo.org" target="_blank">pgrouting-dev@lists.<u></u>osgeo.org</a>>>><br>
            <a href="http://lists.osgeo.org/______mailman/listinfo/pgrouting-dev" target="_blank">http://lists.osgeo.org/______<u></u>mailman/listinfo/pgrouting-dev</a><br>
            <<a href="http://lists.osgeo.org/____mailman/listinfo/pgrouting-dev" target="_blank">http://lists.osgeo.org/____<u></u>mailman/listinfo/pgrouting-dev</a><u></u>><br>
<br>
            <<a href="http://lists.osgeo.org/____mailman/listinfo/pgrouting-dev" target="_blank">http://lists.osgeo.org/____<u></u>mailman/listinfo/pgrouting-dev</a><br>
            <<a href="http://lists.osgeo.org/__mailman/listinfo/pgrouting-dev" target="_blank">http://lists.osgeo.org/__<u></u>mailman/listinfo/pgrouting-dev</a><u></u>>__><br>
<br>
              <<a href="http://lists.osgeo.org/____mailman/listinfo/pgrouting-dev" target="_blank">http://lists.osgeo.org/____<u></u>mailman/listinfo/pgrouting-dev</a><br>
            <<a href="http://lists.osgeo.org/__mailman/listinfo/pgrouting-dev" target="_blank">http://lists.osgeo.org/__<u></u>mailman/listinfo/pgrouting-dev</a><u></u>><br>
<br>
            <<a href="http://lists.osgeo.org/__mailman/listinfo/pgrouting-dev" target="_blank">http://lists.osgeo.org/__<u></u>mailman/listinfo/pgrouting-dev</a><br>
            <<a href="http://lists.osgeo.org/mailman/listinfo/pgrouting-dev" target="_blank">http://lists.osgeo.org/<u></u>mailman/listinfo/pgrouting-dev</a><u></u>>__>__><br>
<br>
<br>
<br>
<br>
                     ______________________________<u></u>_____________________<br>
                     pgrouting-dev mailing list<br>
            <a href="mailto:pgrouting-dev@lists.osgeo.org" target="_blank">pgrouting-dev@lists.osgeo.org</a><br>
            <mailto:<a href="mailto:pgrouting-dev@lists.osgeo.org" target="_blank">pgrouting-dev@lists.<u></u>osgeo.org</a>><br>
            <mailto:<a href="mailto:pgrouting-dev@lists." target="_blank">pgrouting-dev@lists.</a>__<a href="http://osgeo.org" target="_blank"><u></u>osgeo.org</a><br>
            <mailto:<a href="mailto:pgrouting-dev@lists.osgeo.org" target="_blank">pgrouting-dev@lists.<u></u>osgeo.org</a>>><br>
            <a href="http://lists.osgeo.org/____mailman/listinfo/pgrouting-dev" target="_blank">http://lists.osgeo.org/____<u></u>mailman/listinfo/pgrouting-dev</a><br>
            <<a href="http://lists.osgeo.org/__mailman/listinfo/pgrouting-dev" target="_blank">http://lists.osgeo.org/__<u></u>mailman/listinfo/pgrouting-dev</a><u></u>><br>
<br>
            <<a href="http://lists.osgeo.org/__mailman/listinfo/pgrouting-dev" target="_blank">http://lists.osgeo.org/__<u></u>mailman/listinfo/pgrouting-dev</a><br>
            <<a href="http://lists.osgeo.org/mailman/listinfo/pgrouting-dev" target="_blank">http://lists.osgeo.org/<u></u>mailman/listinfo/pgrouting-dev</a><u></u>>__><br>
<br>
<br>
                 ______________________________<u></u>_____________________<br>
                 pgrouting-dev mailing list<br>
            <a href="mailto:pgrouting-dev@lists.osgeo.org" target="_blank">pgrouting-dev@lists.osgeo.org</a><br>
            <mailto:<a href="mailto:pgrouting-dev@lists.osgeo.org" target="_blank">pgrouting-dev@lists.<u></u>osgeo.org</a>><br>
            <mailto:<a href="mailto:pgrouting-dev@lists." target="_blank">pgrouting-dev@lists.</a>__<a href="http://osgeo.org" target="_blank"><u></u>osgeo.org</a><br>
            <mailto:<a href="mailto:pgrouting-dev@lists.osgeo.org" target="_blank">pgrouting-dev@lists.<u></u>osgeo.org</a>>><br>
            <a href="http://lists.osgeo.org/____mailman/listinfo/pgrouting-dev" target="_blank">http://lists.osgeo.org/____<u></u>mailman/listinfo/pgrouting-dev</a><br>
            <<a href="http://lists.osgeo.org/__mailman/listinfo/pgrouting-dev" target="_blank">http://lists.osgeo.org/__<u></u>mailman/listinfo/pgrouting-dev</a><u></u>><br>
<br>
            <<a href="http://lists.osgeo.org/__mailman/listinfo/pgrouting-dev" target="_blank">http://lists.osgeo.org/__<u></u>mailman/listinfo/pgrouting-dev</a><br>
            <<a href="http://lists.osgeo.org/mailman/listinfo/pgrouting-dev" target="_blank">http://lists.osgeo.org/<u></u>mailman/listinfo/pgrouting-dev</a><u></u>>__><br>
<br>
<br>
<br>
<br>
            ______________________________<u></u>___________________<br>
            pgrouting-dev mailing list<br>
            <a href="mailto:pgrouting-dev@lists.osgeo.org" target="_blank">pgrouting-dev@lists.osgeo.org</a><br>
            <mailto:<a href="mailto:pgrouting-dev@lists.osgeo.org" target="_blank">pgrouting-dev@lists.<u></u>osgeo.org</a>><br>
            <a href="http://lists.osgeo.org/__mailman/listinfo/pgrouting-dev" target="_blank">http://lists.osgeo.org/__<u></u>mailman/listinfo/pgrouting-dev</a><br>
            <<a href="http://lists.osgeo.org/mailman/listinfo/pgrouting-dev" target="_blank">http://lists.osgeo.org/<u></u>mailman/listinfo/pgrouting-dev</a><u></u>><br>
<br>
<br>
        ______________________________<u></u>___________________<br>
        pgrouting-dev mailing list<br>
        <a href="mailto:pgrouting-dev@lists.osgeo.org" target="_blank">pgrouting-dev@lists.osgeo.org</a> <mailto:<a href="mailto:pgrouting-dev@lists.osgeo.org" target="_blank">pgrouting-dev@lists.<u></u>osgeo.org</a>><br>

        <a href="http://lists.osgeo.org/__mailman/listinfo/pgrouting-dev" target="_blank">http://lists.osgeo.org/__<u></u>mailman/listinfo/pgrouting-dev</a><br>
        <<a href="http://lists.osgeo.org/mailman/listinfo/pgrouting-dev" target="_blank">http://lists.osgeo.org/<u></u>mailman/listinfo/pgrouting-dev</a><u></u>><br>
<br>
<br>
<br>
<br>
<br>
______________________________<u></u>_________________<br>
pgrouting-dev mailing list<br>
<a href="mailto:pgrouting-dev@lists.osgeo.org" target="_blank">pgrouting-dev@lists.osgeo.org</a><br>
<a href="http://lists.osgeo.org/mailman/listinfo/pgrouting-dev" target="_blank">http://lists.osgeo.org/<u></u>mailman/listinfo/pgrouting-dev</a><br>
<br>
</div></div></blockquote><div class="HOEnZb"><div class="h5">
<br>
______________________________<u></u>_________________<br>
pgrouting-dev mailing list<br>
<a href="mailto:pgrouting-dev@lists.osgeo.org" target="_blank">pgrouting-dev@lists.osgeo.org</a><br>
<a href="http://lists.osgeo.org/mailman/listinfo/pgrouting-dev" target="_blank">http://lists.osgeo.org/<u></u>mailman/listinfo/pgrouting-dev</a><br>
</div></div></blockquote></div><br></div>