[mapserver-dev] FW: [OSGeo-Discuss] OSGeo has been accepted at Google Summer of Code!

[Even Rouault]

Selon Olivier Courtin <olivier.courtin at oslandia.com>:

> WFS and Filter Encoding are really closes to SQL concepts (in fact it's just
> a
> Web Service API in front of a Spatial Database)
>
> So i don't think that using OGR will bring us all the expressivity of plain
> SQL.

I don't see a problem here. OGR justs forwards SQL requests to the RDBMS own SQL
engine when it is available (otherwise OGR has its own generic simplistic SQL
engine, but without any spatial support currently, other than bounding box
intersection through the SetSpatialFilter() API). So if you target OGR backends
that are RDBMS with SQL/MM support, that should be OK. The part that translates
Filter Encoding to SQL/MM would be done in a generic way in TinyOWS.

>
> And also maybe we will loose a bit of performances rather than a direct
> database
> access.

Hum, we are talking about client/servers that interact through the Web, so I
doubt the OGR overhead makes a significant difference.

>
> But yes it imply to do quite the same job (again) with Oracle Spatial sooner
> or later...
>
> --
> Olivier
>
>
>
>
>
>
>
>
> _______________________________________________
> mapserver-dev mailing list
> mapserver-dev at lists.osgeo.org
> http://lists.osgeo.org/mailman/listinfo/mapserver-dev
>