[OSGeo-Discuss] Raster data on a DBMS

P Kishor punk.kish at gmail.com
Mon Nov 3 20:08:13 PST 2008


On 11/3/08, Gilberto Camara <gilberto.camara at inpe.br> wrote:
..
>
>  Allow me to reiterate my earlier argument, which is
>  that FOSS4G should **allow** users the option of storing
>  raster data in a database. Storing images in a database
>  is not recommended in each and every situation.
>  The user should have the option, according to his needs.
>
..

There are two kinds of users in this world when it comes to storing
raster data... and, fortunately, the world is big enough for both of
them.

The FOSS in FOSS4G gives each kind to develop what they need/want.
There is no concept of "allowing" anyone the option.

Gilberto, my sense is that instead of convincing someone else that
storing rasters in a db can/may be a better option in some cases,
energy and resources might be better deployed in making options
available, and making existing options easier and more powerful.

In another thread you wrote --

"INPE┬┤s FOSS4G developement of raster data on RDBMS using the
TerraLib library is a tangible proof of concept. TerraAmazon
(built using TerraLib) is INPE's OS solution for monitoring
tropical forests operationally."

While I have an earlier version of TerraLib/TerraView that I got when
I visited INPE last year, I am not familiar with its raster-in-db
capabilities. Making and distributing a product that can work on
different platforms, Windows, Macs, *nix, will be the best marketing
for such capabilities, and provide the options that you advocate.

Fwiw, O'Reilly's database war stories series has an article on Flickr
that store metadata in a db and the photos themselves on directly in
the filesystem. See
<http://radar.oreilly.com/archives/2006/04/database_war_stories_3_flickr.html>.
Other stories in that series are also very instructive.

Apple's Aperture does the same thing... all the metadata are in SQLite
(Coredata on Mac OS X) while the images themselves are stored in
folders on the disk -- albeit hidden from the user. See
<http://www.bagelturf.com/aparticles/tips/tiplib/index.php> and
<http://www.mungosmash.com/archives/2005/12/the_aperture_da.php>.

Of course, both of the above examples are of the "high number of small
size images" variety while the raster/geographic applications are of
the "relatively fewer number of very large size images" kind. Still,
the case studies are instructive.

I think the best argument for any approach is the approach itself. The
free and open part in FOSS4G allows any and all approaches to exist.
Scratch your own itch, meet your own need, then put it out for others
to use. If they like it, they will adopt it and improve it. If they
don't like it, it will die a timely death. Darwinian at its best.


-- 
Puneet Kishor http://punkish.eidesis.org/
Nelson Institute for Environmental Studies http://www.nelson.wisc.edu/
Open Source Geospatial Foundation (OSGeo) http://www.osgeo.org/



More information about the Discuss mailing list