Hm, I am just writing a proposal for a large German/Dutch project on
historical mapping in the Euregio (Maastricht/Aachen/Liège region), in
the order of €1000000. It will be hosted at SARA Computing Center of the
University of Amsterdam, with 5 dedicated nodes within a large Linux
cluster, and completely based on Open Source (MapServer, PostGIS, GDAL
and PL/R). Until now, I had not thought of GRASS for this project,
although I have been working with GRASS from the early nineties, as my
project is essentially directed towards end-users, and as you probably
know, it is horribly difficult to get this kind of people working with
GRASS. Things are getting better nowadays with programs like QGis and
the interactive interface, but it is still a long haul.
Technically and financially, it would be easy to extend the number of
nodes to allow research on parallellizing Open Source, but I do not have
the time and knowledge to coordinate it (my main interest and expertise
is in historical cartography). I have been thinking of contracting the
main developers of MapServer etc, whom I know reasonably well, to see if
they could organise something along those lines, e.g. within the OsGEO
communities. Again, the problem is not so much finance, but getting
people who can invest their time and knowledge to formulate specific
goals, and coordinate and guide the process so as to deliver concrete
results over a period of (say) three years. This has to be organised in
some formal way, as it concerns quite a lot of money. Any interest for
this with the GRASS-developers?
Jan
Dr. J. Hartmann
Department of Geography
University of Amsterdam
Wolf Bergenheim wrote:
On 09.07.2007 14:57, Jan Hartmann wrote:
I would like to add an important aspect I am working on now: cluster
computing. There is a big potential in parallellizing algorithms within
GRASS, and that would really be unaffordable with commercial
GIS-software. A Linux cluster nowadays doesn't cost anything any more,
and with free software the possibilities for development are limitless.
That would make an excellent topic for next year's Google Summer of
Code. To parallelisize the raster modules to begin with. Thanks for the
idea :D
This year we got the most interest in task involving algorithms. I hope
this would also be a popular subject :)
--Wolf
_______________________________________________
grassuser mailing list
[email protected]
http://grass.itc.it/mailman/listinfo/grassuser
