Hi all, I'm working on the GDAL Google Maps Engine (GME) driver over then next few months for a school project. I've set up a wiki page for it where you can track my progress. It's here:
http://trac.osgeo.org/gdal/wiki/GMEDriver I'm picking up from where Frank and I left off a few months ago. For now we have a mapping that mimics the Fusion Tables setup, which is that each OGRGMEDataSource corresponds to a single session, and each OGRGMELayer corresponds to a table. This works fine, but GME can do more! :) GME has a concept of a map which contains layers, each layer has a 1:1 relationship with a vector table. Now my proposal is to extend the previous model so that if the user of the library specifies a map instead of a table, we change the mapping so that the OGRGMEDatasource maps to the map and contains several OGRGMELayers which each correspond to each table. Then I would move the connection logic out of OGRGMEDatasource to OGRGMEDriver, which is currently basically only boilerplate code. Is there a precedence for this type of configuration in any other driver? What do you think would this kind of added mapping be worth the extra effort? I'm trying to make the classes as lean as possible, because I might have a need to pull out all the map/layer handling code so that I can re-use it later for another driver. Also I think that the connection management / request throttling code could be usable elsewhere. That's it for now. Let me know if you have any specific wishes regarding the GME driver. --Wolf
_______________________________________________ gdal-dev mailing list [email protected] http://lists.osgeo.org/mailman/listinfo/gdal-dev
