On 1/25/07, Andrea Antonello <andrea.antonello@anonymised.com> wrote:
Hi Randy,
I am really not enough into the specifications issue, so I hope I don't
say complete rubbish.
> I have a question regarding Subject: Re: [udig-devel] JGrass and
> uDig.
>
> I can see where uDig could benefit from JGrass/Grass/OSSIM analysis
> functions on the client side after building an image of interest. OpenGL
> wrapper JSR2(231) makes sense for high performance GPU utilization on the
> client (Vista is bound to bump GPU availability onto the mainstream over the
> next year or so) but I'm wondering if server side functions would also be of
> use.
It depends really. I am not super super expert but I have done much
experiments with new OPenGL pipeline (not in the last months to be
honest).
There is a distinction to be made, using opengl for doing rendering
when you have a good gpu is great but using opengl for doing image
processing may sometime be not optima because you might be end up
moving your data between the graphic card memory and the main memory.
When data is quite huge and you do extensive tiling using opengl might
end up in lower performances.
In my opinion opengl is good for visualization purposes with a good
graphic card not (in general) for processing.
To be honest, I have no answer for that, never thought about it.
> I'm thinking about the OGC Grid coverage spec with its
> 2.5 GP_GridAnalysis (optional)
> 3.1.5 GP_GridCoverageProcessor
> 3.1.8.1 GP_Operation (optional)
>
> JGrass contains a wealth of raster operations! Could these become
> part of grid coverage (WCS) GP_GridAnaylsis and GP_Operations in the
> Geoserver roadmap?
I would split this in two steps:
1>Geotools processing capabilities (Martin help! 
)
Geotools already has a GridProcessing package which of course could be
improved (well actually it is under improvement). Adding operations
based on JGrass should not be that difficult and could be quite
interesting.
2>Server side I am looking with interests in the GRASS-WPS experiments
and I would really like to give it a try. It would anyway interesting
to set up some sort of WPS using some JGrass capabilities once we hide
them under geotools. It would be great to set up some sort of WCPS for
doing some classic raster operations like classifications nd the like.
Don't know about geoserver plans, but with JGrass we wanted at some
point create the possibility to follow the WPS standard, which I thought
to be what you are talking about? I'm a bit confused.
> The possibility of server side chaining on coverage operations is also
> interesting.
I agree, remember that WCPS is WPS+WCS and it is simply a best practice paper.
I have not had much time yet to look into this JGrass thread,
hopefully tomorrow I will start reading all the passed emails :-(.
Thx,
Simone.
Yes, the concept of linkable command could make a big sense serverside,
since also you bypass the creation of temporary maps, which otherwise
slow down your process.
Andrea
_______________________________________________
User-friendly Desktop Internet GIS (uDig)
http://udig.refractions.net
http://lists.refractions.net/mailman/listinfo/udig-devel
--
-------------------------------------------------------
Eng. Simone Giannecchini
President /CEO GeoSolutions
http://www.geo-solutions.it
-------------------------------------------------------