Raster Analysis and MapServer

[Blammo]

A thought, what about exposing the tool as a Web Service, similar call 
to MapServer, only via a WebService.  This would help out with the 
Thread/Safe issue too as Apache keeps this things generally under 
control.  Large processes that take some time, might be a problem for 
Timeouts.  It's just a thought.

I actually thought about this last year when we were first starting out 
with our Services architecture at the City.



bobb

Lowell.Filak wrote:

>Martin, Aaron, & others. 
>
>This is good news.
>I have been trying to get processes like this documented to the wiki (before 
>the spamming) and to the new site by replying to emails such as these.
>I finally gave up in favor of hiring a contractor to develop a (hopefully) 
>uniform scripting language interface between MapScript & GRASS. As a result 
>please expect to be contacted by the aforementioned contractor during the 
>upcoming feasibility study. 
>
>Lowell 
>
>Martin Hoegh writes: 
>
>  
>
>>Hi,
>>I also use Grass for backend. I've written a python script using pexpect to
>>start and "drive" a grass session. In this way I can use grass's modules
>>through the script as if I was sitting in front of the terminal. The python
>>script exposes its functions via SOAP, which allows me to call the script
>>with what ever I want - normally php. The down side is that grass is not
>>thread safety, which means you can't start more than one session in a mapset
>>at the time. My script waits in a loop until the previous spawned session
>>ends. This is not good for heavy traffic! After calling grass you can use
>>MapServer to render and display the map, neither directly from the grass
>>database or let grass export the data to a image file. 
>>
>>Martin
>>    
>>
> 
>
>
>  
>

-------------- next part --------------
An HTML attachment was scrubbed...
URL: <http://lists.osgeo.org/pipermail/mapserver-users/attachments/20051222/c7cd9ce6/attachment.htm>