Awesome,

once again i'm surprised in a good way by whats possible in gs and its 
friends.

it looks like the rendering side is very advanced here, but on the other 
side i really want to do as little pre-processing as possible, as our 
data is volatile, hence using the netcdf directly by some method 
(perhaps using the wms 3 time filters ?) is interesting, i will follow 
up on this with some of you.

i know that gdal supports netcdf to some degree, but i'm not sure if 
geoserver can be configured to render a particular variable / dimension 
slice using that.

cheers for all your excellent advice

-i



Le 26/03/2011 9:32 AM, Andrea Aime a écrit :
> On Sat, Mar 26, 2011 at 9:13 AM, Luca Morandini<lmorand...@ieee.org>  wrote:
>> Oops, it appears I made a sily question... actually, I failed to notice 
>> the<<pipe
>> a process (yes, a WPS one) inside the SLD>>  sentence, which means it is 
>> *very*
>> pluggable.
>>
>> This mechanism takes the separation of content from presentation to a new 
>> level,
>> which I like immensely... but I cannot help thinking that, performance-wise, 
>> the
>> WPS overhead needed for every GetMap call is not very reassuring.
>>
>> Anyway, you're the expert on GS performance here.
> Indeed the processes you've seen have been fine tuned to work at best
> performance,
> you cannot take a random WPS process and pretend it will be worthy of usage in
> a on the fly rendering transformation.
>
> Then again, you have a choice:
> - preprocess the data, taking the extra effort of orchestrating the
> data ingestion,
>    transformation in the various representations needed, publishing of
> the various
>    layers in GS with styles and all, and long term storage of that somewhat
>    redundant data or
> - on the fly transformation, which also allows you to pass dynamically the
>    transformation params from the client (see sld param substitution, another
>    custom GS extension), where you just register one layer, and you have to
>    be very careful at optimizing the transformation processes
>
> For the specific case we developed this feature for the choice was a no 
> brainer,
> when you have 100k layers and the viewer wants to specify the isoline levels
> inside the client, for that specific session only, the first approach just 
> does
> not really work well
>
> Cheers
> Andrea
>
>


------------------------------------------------------------------------------
Enable your software for Intel(R) Active Management Technology to meet the
growing manageability and security demands of your customers. Businesses
are taking advantage of Intel(R) vPro (TM) technology - will your software 
be a part of the solution? Download the Intel(R) Manageability Checker 
today! http://p.sf.net/sfu/intel-dev2devmar
_______________________________________________
Geoserver-users mailing list
Geoserver-users@lists.sourceforge.net
https://lists.sourceforge.net/lists/listinfo/geoserver-users

Reply via email to