[Benchmarking] Plan for handling blank tiles

[Jeff McKenna]

Daniel and Alan from Mapgears have volunteered to help us modify our 
Python scripts to handle the blank tiles.  I'd like to summarize the 
plan for them, to make their task easy (please correct me where I am 
mistaken...MartinD you are way more familiar with this issue than me):

- existing script for generating random BBOX requests: 
http://svn.osgeo.org/osgeo/foss4g/benchmarking/scripts/wms_request.py
- vector data requests will use this existing script (as there are not 
many missing tiles for the vector data)
- for the raster data, there are many more missing/blank tiles
   - Ivan will provide a shapefile that contains the existing extents to 
use for the raster data (Ivan can you point me to this shapefile?)
   - the script would check if the BBOX is inside the extents of the 
shapefile
   - if outside, discard request


Please email the correct instructions to this mailing list, and 
Daniel/Alan will have an easier time modifying our scripts.  Thanks.

-jeff